# Claude Fable 5Max

> Source: Benchize (https://benchize.com) — https://benchize.com/models/claude-fable-5-max. When citing or reusing this content, credit "Benchize" and link to https://benchize.com/models/claude-fable-5-max.

Anthropic

Canonical URL: https://benchize.com/models/claude-fable-5-max
Last updated: Wed, 21 Oct 2026 18:05:10 GMT

Reproduction and quotation must name Benchize, link to the canonical HTML URL, and retain individual source citations.

Anthropic

## Benchmarks

| Benchmark | Score | ± | Rank | Cost | Latency | Source |  |
| --- | --- | --- | --- | --- | --- | --- | --- |
| [Harvey's Legal Agent Benchmark v1 / criteria pass rate insurance](https://benchize.com/benchmarks/vals-ai%3Ahlab-v1%3Acriteria_pass_rate__insurance.md) | 93.41% | SE ±1.51 pp | 2 / 49 | $15.61 | 1,417.02source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 93.407 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 15.611186 \| USD \| per_test \| Lower is better \| \| `latency` \| 1417.02 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.51 pp sourcePopulation 81 stderr 1.505 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__insurance/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [MMLU Pro / chemistry](https://benchize.com/benchmarks/vals-ai%3Ammlu_pro%3Achemistry.md) | 93.91% | SE ±0.71 pp | 1 / 40 | $0.09 | 39.62source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_mmlu_pro.y416pXDi.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 93.905 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.094911 \| USD \| per_test \| Lower is better \| \| `latency` \| 39.622 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.71 pp sourcePopulation 138 stderr 0.711 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_mmlu_pro.y416pXDi.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:46:06.503Z Evidence SHA-256 `dab025507f46c716dd8d145e90ec0a737d9c56219853f41e4b1cc3c5e8c222b4` Source locator `/tasks/chemistry/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [MedCode / overall](https://benchize.com/benchmarks/vals-ai%3Amedcode%3Aoverall.md) | 56.07% | SE ±2.2 pp | 4 / 49 | $0.59 | 91.44source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_medcode.DehpEDgG.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 56.07 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.591071 \| USD \| per_test \| Lower is better \| \| `latency` \| 91.439 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 30000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±2.2 pp sourcePopulation 107 stderr 2.203 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_medcode.DehpEDgG.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:59.516Z Evidence SHA-256 `988a6ae1e07e33ce031572ed4a804be11c36ee97a1dd75cbe4740fff3b534f8c` Source locator `/tasks/overall/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [CharXiv (RQ)](https://benchize.com/benchmarks/charxiv-rq.md) | 88.9% | — | 1 / 5 | — | — | [Moonshot AI · sha256:0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c](https://www.kimi.ai/blog/kimi-k3) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 88.9 \| percent \| reported \| Higher is better \| ### Configuration Fallback policy unspecified fallbackEnabled true Reasoning effort max runs 3 sourceFallbackLabel with fallback subset RQ thinkingEnabled true toolsEnabled false Fallback is enabled for refusals; this does not mean every task used a fallback model. ### Reported statistics runCount 3 ### Source Score publisher [Moonshot AI](https://www.kimi.ai/blog/kimi-k3) Collected via [moonshot k3](https://www.kimi.ai/blog/kimi-k3) Publisher reference Referenced by the collection source; the upstream result has not been independently captured. Benchmark [https://www.kimi.ai/blog/kimi-k3](https://www.kimi.ai/blog/kimi-k3) [Collected via](https://www.kimi.ai/blog/kimi-k3) Source Moonshot AI Original model label Claude Fable 5 (max, with fallback) Source version sha256:0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c Observed 2026-10-21T16:31:44.130Z Evidence SHA-256 `0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c` Source locator `/flight/sheet/data/25/scores/1` |  |  |  |  |  |  |  |
| [Terminal-Bench-Science / overall](https://benchize.com/benchmarks/vals-ai%3Aterminal-bench-science%3Aoverall.md) | 15.71% | SE ±4.38 pp | 11 / 30 | $51.99 | 7,157.67source duration | [Vals AI · 0.1](https://www.vals.ai/_astro/benchmark_view_terminal_bench_science.o-HR8Nyt.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 15.714 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 51.990885 \| USD \| per_test \| Lower is better \| \| `latency` \| 7157.669 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Mini-SWE-agent maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±4.38 pp sourcePopulation 41 stderr 4.381 tokenTotals {"cache_read_tokens":2569404286,"cache_write_tokens":30031671,"input_tokens":5198895542,"output_tokens":27449950,"reasoning_tokens":10535367} ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_terminal_bench_science.o-HR8Nyt.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 0.1 Observed 2026-10-21T16:46:03.529Z Evidence SHA-256 `a30a8618deaecf018ea1c8c9f675576cdd9ab79742d05788067db49afcc5f91b` Source locator `/tasks/overall/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [MMLU Pro / philosophy](https://benchize.com/benchmarks/vals-ai%3Ammlu_pro%3Aphilosophy.md) | 90.58% | SE ±1.31 pp | 1 / 40 | $0.07 | 24.04source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_mmlu_pro.y416pXDi.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 90.581 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.069071 \| USD \| per_test \| Lower is better \| \| `latency` \| 24.035 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.31 pp sourcePopulation 138 stderr 1.308 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_mmlu_pro.y416pXDi.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:46:06.503Z Evidence SHA-256 `dab025507f46c716dd8d145e90ec0a737d9c56219853f41e4b1cc3c5e8c222b4` Source locator `/tasks/philosophy/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [Harvey's Legal Agent Benchmark v1 / intellectual property](https://benchize.com/benchmarks/vals-ai%3Ahlab-v1%3Aintellectual_property.md) | 0% | SE ±0 pp | 6 / 49 | $19.84 | 1,536.33source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 0 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 19.843555 \| USD \| per_test \| Lower is better \| \| `latency` \| 1536.328 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0 pp sourcePopulation 81 stderr 0 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/intellectual_property/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [Code Migration / overall](https://benchize.com/benchmarks/vals-ai%3Acode-migration%3Aoverall.md) | 55.07% | SE ±4.61 pp | 9 / 47 | $112.1 | 6,559.39source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_code_migration.Dttzoqk3.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 55.065 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 112.098109 \| USD \| per_test \| Lower is better \| \| `latency` \| 6559.388 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±4.61 pp sourcePopulation 79 stderr 4.605 tokenTotals {"cache_read_tokens":13465564599,"cache_write_tokens":84247569,"input_tokens":13549878096,"output_tokens":49676333,"reasoning_tokens":null} ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_code_migration.Dttzoqk3.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:46:12.712Z Evidence SHA-256 `99194597f2c76c402195be47839409046e18bd30454616da6dd40b736a088fa4` Source locator `/tasks/overall/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [LiveCodeBench / overall](https://benchize.com/benchmarks/vals-ai%3Alcb%3Aoverall.md) | 89.78% | SE ±0.89 pp | 2 / 43 | $0.43 | 118.53source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_lcb.miyRUvVa.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 89.778 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.42816 \| USD \| per_test \| Lower is better \| \| `latency` \| 118.53 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.89 pp sourcePopulation 143 stderr 0.892 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_lcb.miyRUvVa.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:46:14.220Z Evidence SHA-256 `1a7c7dabc9d95a0182886b1edc42623987cac2095a0d407a5e334f37acdfa2f5` Source locator `/tasks/overall/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [MMLU Pro / math](https://benchize.com/benchmarks/vals-ai%3Ammlu_pro%3Amath.md) | 97.56% | SE ±0.42 pp | 1 / 40 | $0.08 | 21.37source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_mmlu_pro.y416pXDi.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 97.557 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.07686 \| USD \| per_test \| Lower is better \| \| `latency` \| 21.371 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.42 pp sourcePopulation 138 stderr 0.42 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_mmlu_pro.y416pXDi.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:46:06.503Z Evidence SHA-256 `dab025507f46c716dd8d145e90ec0a737d9c56219853f41e4b1cc3c5e8c222b4` Source locator `/tasks/math/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [Tax Agent Bench / weighted pass numbers calculations](https://benchize.com/benchmarks/vals-ai%3Atax_agent_bench%3Aweighted_pass__numbers_calculations.md) | 83.38% | SE ±6.17 pp | 7 / 46 | $2.3 | 325.85source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_tax_agent_bench.BFpexnpS.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 83.382 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 2.303921 \| USD \| per_test \| Lower is better \| \| `latency` \| 325.851 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±6.17 pp sourcePopulation 73 stderr 6.171 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_tax_agent_bench.BFpexnpS.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:55.958Z Evidence SHA-256 `446db182b67efb7923b3187a4040ec3a3ea492b8ad699a2b6d708d714cc063ac` Source locator `/tasks/weighted_pass__numbers_calculations/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [DeepSearchQA (f1-score)](https://benchize.com/benchmarks/deepsearchqa.md) | 94.2% | — | 2 / 3 | — | — | [Moonshot AI · sha256:0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c](https://www.kimi.ai/blog/kimi-k3) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `f1` \| 94.2 \| percent \| reported \| Higher is better \| ### Configuration Fallback policy unspecified fallbackEnabled true Reasoning effort max sourceFallbackLabel with fallback thinkingEnabled true Fallback is enabled for refusals; this does not mean every task used a fallback model. ### Source Score publisher [Moonshot AI](https://www.kimi.ai/blog/kimi-k3) Collected via [moonshot k3](https://www.kimi.ai/blog/kimi-k3) Publisher reference Referenced by the collection source; the upstream result has not been independently captured. Benchmark [https://www.kimi.ai/blog/kimi-k3](https://www.kimi.ai/blog/kimi-k3) [Collected via](https://www.kimi.ai/blog/kimi-k3) Source Moonshot AI Original model label Claude Fable 5 (max, with fallback) Source version sha256:0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c Observed 2026-10-21T16:31:44.130Z Evidence SHA-256 `0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c` Source locator `/flight/sheet/data/10/scores/1` |  |  |  |  |  |  |  |
| [Harvey's Legal Agent Benchmark v1 / corporate governance](https://benchize.com/benchmarks/vals-ai%3Ahlab-v1%3Acorporate_governance.md) | 20% | SE ±20 pp | 3 / 49 | $18.66 | 1,644.85source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 20 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 18.661957 \| USD \| per_test \| Lower is better \| \| `latency` \| 1644.848 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±20 pp sourcePopulation 81 stderr 20 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/corporate_governance/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [Terminal-Bench-Science / mathematical sciences](https://benchize.com/benchmarks/vals-ai%3Aterminal-bench-science%3Amathematical_sciences.md) | 11.77% | SE ±8.06 pp | 10 / 30 | — | — | [Vals AI · 0.1](https://www.vals.ai/_astro/benchmark_view_terminal_bench_science.o-HR8Nyt.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 11.765 \| percent \| reported \| Higher is better \| ### Configuration computeEffort max harness Mini-SWE-agent maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±8.06 pp sourcePopulation 41 stderr 8.055 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_terminal_bench_science.o-HR8Nyt.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 0.1 Observed 2026-10-21T16:46:03.529Z Evidence SHA-256 `a30a8618deaecf018ea1c8c9f675576cdd9ab79742d05788067db49afcc5f91b` Source locator `/tasks/mathematical_sciences/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [ARC-AGI-2](https://benchize.com/benchmarks/arc-prize%3Av2_Semi_Private.md) | 89.17% | — | 18 / 212 | 5.45usd per task | — | [arc prize · v2](https://arcprize.org/media/data/leaderboard/v2.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 0.8916666666666667 \| score_fraction \| v2_Semi_Private \| Higher is better \| \| `cost_per_task` \| 5.449076000000001 \| usd_per_task \| v2_Semi_Private \| Lower is better \| ### Configuration evaluationSplit v2_Semi_Private harness Not reported modelGroup anthropic-claude-fable-5 modelType CoT Reasoning effort max sourceModelVersion anthropic-claude-fable-5-max thinkingEnabled true ### Source [Original source](https://arcprize.org/media/data/leaderboard/v2.json) Source arc prize Original model label Claude Fable 5 (Max) Source version v2 Observed 2026-10-21T16:34:20.750Z Evidence SHA-256 `bca68e12cc9a01954ae21b641844c1ecbabe146cb24b87f0f8e9b1db27433c29` Source locator `/evaluations/187` |  |  |  |  |  |  |  |
| [MortgageTax / semantic extraction](https://benchize.com/benchmarks/vals-ai%3Amortgage_tax%3Asemantic_extraction.md) | 61.13% | SE ±1.37 pp | 9 / 31 | $0.17 | 14.99source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_mortgage_tax.DCd-oO5Y.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 61.129 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.166658 \| USD \| per_test \| Lower is better \| \| `latency` \| 14.993 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.37 pp sourcePopulation 98 stderr 1.374 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_mortgage_tax.DCd-oO5Y.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:58.500Z Evidence SHA-256 `e5c844561f97d119d66d70269e726ecfb68be79eae39acf14ead3e95422e648c` Source locator `/tasks/semantic_extraction/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [MMLU Pro / health](https://benchize.com/benchmarks/vals-ai%3Ammlu_pro%3Ahealth.md) | 84.6% | SE ±1.26 pp | 6 / 40 | $0.08 | 20.08source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_mmlu_pro.y416pXDi.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 84.597 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.084721 \| USD \| per_test \| Lower is better \| \| `latency` \| 20.082 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.26 pp sourcePopulation 138 stderr 1.262 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_mmlu_pro.y416pXDi.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:46:06.503Z Evidence SHA-256 `dab025507f46c716dd8d145e90ec0a737d9c56219853f41e4b1cc3c5e8c222b4` Source locator `/tasks/health/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [Legal Research Bench / weighted pass business commercial](https://benchize.com/benchmarks/vals-ai%3Alegal_research%3Aweighted_pass__business_commercial.md) | 91.37% | SE ±1.91 pp | 4 / 50 | $9.52 | 1,482.1source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_research.DiuL0R8P.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 91.374 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 9.515091 \| USD \| per_test \| Lower is better \| \| `latency` \| 1482.098 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.91 pp sourcePopulation 82 stderr 1.907 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_research.DiuL0R8P.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:52.389Z Evidence SHA-256 `26c58dfd47d0d2dc86f33c6d056f30309e0ee63570e7cd752c593b385f9d862e` Source locator `/tasks/weighted_pass__business_commercial/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [Finance Agent (v2) / all pass](https://benchize.com/benchmarks/vals-ai%3Afabv2%3Aall_pass.md) | 45.69% | SE ±0.87 pp | 14 / 50 | $8.06 | 610.95source duration | [Vals AI · 2](https://www.vals.ai/_astro/benchmark_view_finance_agent_v2.DypCOnwd.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 45.694 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 8.061563 \| USD \| per_test \| Lower is better \| \| `latency` \| 610.952 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.87 pp sourcePopulation 80 stderr 0.871 tokenTotals {"cache_read_tokens":110357130,"cache_write_tokens":201446408,"input_tokens":311882539,"output_tokens":19969523,"reasoning_tokens":null} ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_finance_agent_v2.DypCOnwd.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 2 Observed 2026-10-21T16:45:57.035Z Evidence SHA-256 `f731b20a26c126725d125519f2b5841aa45254656dba859a9570d7ce7e5d8191` Source locator `/tasks/all_pass/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [FrontierSWE](https://benchize.com/benchmarks/frontierswe%3Adominance.md) | 86.6% | — | 1 / 6 | — | — | [Moonshot AI · sha256:0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c](https://www.kimi.ai/blog/kimi-k3) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `dominance` \| 86.6 \| percent \| dominance_2026-07-16 \| Higher is better \| ### Configuration Fallback policy unspecified fallbackEnabled true normalization dominance normalizationDate 2026-07-16 Reasoning effort max sourceFallbackLabel with fallback thinkingEnabled true Fallback is enabled for refusals; this does not mean every task used a fallback model. ### Source Score publisher [Moonshot AI](https://www.kimi.ai/blog/kimi-k3) Collected via [moonshot k3](https://www.kimi.ai/blog/kimi-k3) Publisher reference Referenced by the collection source; the upstream result has not been independently captured. Benchmark [https://www.kimi.ai/blog/kimi-k3](https://www.kimi.ai/blog/kimi-k3) [Collected via](https://www.kimi.ai/blog/kimi-k3) Source Moonshot AI Original model label Claude Fable 5 (max, with fallback) Source version sha256:0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c Observed 2026-10-21T16:31:44.130Z Evidence SHA-256 `0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c` Source locator `/flight/sheet/data/3/scores/1` |  |  |  |  |  |  |  |
| [Legal Research Bench / weighted pass family](https://benchize.com/benchmarks/vals-ai%3Alegal_research%3Aweighted_pass__family.md) | 77.9% | SE ±5.85 pp | 7 / 50 | $11.01 | 1,643.01source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_research.DiuL0R8P.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 77.897 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 11.011763 \| USD \| per_test \| Lower is better \| \| `latency` \| 1643.011 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±5.85 pp sourcePopulation 82 stderr 5.851 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_research.DiuL0R8P.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:52.389Z Evidence SHA-256 `26c58dfd47d0d2dc86f33c6d056f30309e0ee63570e7cd752c593b385f9d862e` Source locator `/tasks/weighted_pass__family/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [ArXivLean 06/2026](https://benchize.com/benchmarks/matharena%3Aarxivlean--june.md) | 15.22% | 95% CI 4.84–25.6% | 7 / 10 | 8.44usd per problem | — | [matharena · arxivlean--june](https://matharena.ai/competition_tables/arxivlean--june) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 15.22 \| percent \| competition \| Higher is better \| \| `cost_per_problem` \| 8.44 \| usd_per_problem \| competition \| Lower is better \| ### Configuration competition arxivlean--june predicted false Reasoning effort max sourceModelName Claude-Fable-5 (max) sourceModelVersion anthropic/fable_5_max_128 thinkingEnabled true ### Reported statistics Uncertainty 95% confidence interval: 4.84–25.6% confidenceInterval95 10.38 problems 46 ### Source [Original source](https://matharena.ai/competition_tables/arxivlean--june) Source matharena Original model label Claude-Fable-5 (max) Source version arxivlean--june Observed 2026-10-21T16:52:22.483Z Evidence SHA-256 `0869c2556b20f1879eb0ea99ea8151e805471cd9d6be57c652f764a775491b7d` Source locator `/table/rows/10` |  |  |  |  |  |  |  |
| [LegalBench / overall](https://benchize.com/benchmarks/vals-ai%3Alegal_bench%3Aoverall.md) | 88.56% | SE ±0.33 pp | 1 / 44 | $0.03 | 8.96source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 88.561 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.030639 \| USD \| per_test \| Lower is better \| \| `latency` \| 8.964 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.33 pp sourcePopulation 151 stderr 0.331 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/overall/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [Harvey's Legal Agent Benchmark v1 / emerging companies venture capital](https://benchize.com/benchmarks/vals-ai%3Ahlab-v1%3Aemerging_companies_venture_capital.md) | 20% | SE ±20 pp | 6 / 49 | $17.39 | 1,722.23source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 20 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 17.394862 \| USD \| per_test \| Lower is better \| \| `latency` \| 1722.228 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±20 pp sourcePopulation 81 stderr 20 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/emerging_companies_venture_capital/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [MLS Bench](https://benchize.com/benchmarks/mls-bench-lite.md) | 49.9% | — | 1 / 6 | — | — | [Moonshot AI · sha256:0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c](https://www.kimi.ai/blog/kimi-k3) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 49.9 \| percent \| reported \| Higher is better \| ### Configuration Fallback policy unspecified fallbackEnabled true harness Claude Code Reasoning effort max sourceFallbackLabel with fallback thinkingEnabled true Fallback is enabled for refusals; this does not mean every task used a fallback model. ### Source Score publisher [Moonshot AI](https://www.kimi.ai/blog/kimi-k3) Collected via [moonshot k3](https://www.kimi.ai/blog/kimi-k3) Publisher reference Referenced by the collection source; the upstream result has not been independently captured. Benchmark [https://www.kimi.ai/blog/kimi-k3](https://www.kimi.ai/blog/kimi-k3) [Collected via](https://www.kimi.ai/blog/kimi-k3) Source Moonshot AI Original model label Claude Fable 5 (max, with fallback) Source version sha256:0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c Observed 2026-10-21T16:31:44.130Z Evidence SHA-256 `0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c` Source locator `/flight/sheet/data/6/scores/1` |  |  |  |  |  |  |  |
| [MortgageTax / numerical extraction](https://benchize.com/benchmarks/vals-ai%3Amortgage_tax%3Anumerical_extraction.md) | 76.71% | SE ±1.19 pp | 11 / 31 | $0.18 | 17.2source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_mortgage_tax.DCd-oO5Y.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 76.709 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.183368 \| USD \| per_test \| Lower is better \| \| `latency` \| 17.2 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.19 pp sourcePopulation 98 stderr 1.192 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_mortgage_tax.DCd-oO5Y.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:58.500Z Evidence SHA-256 `e5c844561f97d119d66d70269e726ecfb68be79eae39acf14ead3e95422e648c` Source locator `/tasks/numerical_extraction/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [CorpFin v2 / max fitting context](https://benchize.com/benchmarks/vals-ai%3Acorp_fin_v2%3Amax_fitting_context.md) | 77.04% | SE ±1.44 pp | 1 / 39 | $3.73 | 81.47source duration | [Vals AI · 2](https://www.vals.ai/_astro/benchmark_view_corp_fin_v2.CectKwl1.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 77.04 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 3.734685 \| USD \| per_test \| Lower is better \| \| `latency` \| 81.472 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.44 pp sourcePopulation 134 stderr 1.436 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_corp_fin_v2.CectKwl1.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 2 Observed 2026-10-21T16:45:57.998Z Evidence SHA-256 `7cf2217ab3728d5a84bffce82aab9fc52d7f810b0cbe8c65b387df47f3d31c9b` Source locator `/tasks/max_fitting_context/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [Harvey's Legal Agent Benchmark v1 / criteria pass rate banking finance](https://benchize.com/benchmarks/vals-ai%3Ahlab-v1%3Acriteria_pass_rate__banking_finance.md) | 92.57% | SE ±1.6 pp | 7 / 49 | $15.8 | 1,390.37source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 92.565 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 15.803642 \| USD \| per_test \| Lower is better \| \| `latency` \| 1390.374 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.6 pp sourcePopulation 81 stderr 1.602 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__banking_finance/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [Harvey's Legal Agent Benchmark v1 / tax](https://benchize.com/benchmarks/vals-ai%3Ahlab-v1%3Atax.md) | 0% | SE ±0 pp | 3 / 49 | $17.08 | 1,763.37source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 0 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 17.075748 \| USD \| per_test \| Lower is better \| \| `latency` \| 1763.37 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0 pp sourcePopulation 81 stderr 0 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/tax/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [TaxEval v2 / stepwise reasoning](https://benchize.com/benchmarks/vals-ai%3Atax_eval_v2%3Astepwise_reasoning.md) | 88.06% | SE ±0.93 pp | 5 / 40 | $0.22 | 56.87source duration | [Vals AI · 2](https://www.vals.ai/_astro/benchmark_view_tax_eval_v2.B3yoffr8.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 88.062 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.220133 \| USD \| per_test \| Lower is better \| \| `latency` \| 56.865 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.93 pp sourcePopulation 145 stderr 0.927 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_tax_eval_v2.B3yoffr8.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 2 Observed 2026-10-21T16:45:59.000Z Evidence SHA-256 `8b354a6ae848f5bb25a2e8b061b5f576d5b9e0c11b941dc1e6bed078e6529cd6` Source locator `/tasks/stepwise_reasoning/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [Harvey's Legal Agent Benchmark v1 / criteria pass rate emerging companies venture capital](https://benchize.com/benchmarks/vals-ai%3Ahlab-v1%3Acriteria_pass_rate__emerging_companies_venture_capital.md) | 93.86% | SE ±1.41 pp | 7 / 49 | $17.39 | 1,722.23source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 93.857 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 17.394862 \| USD \| per_test \| Lower is better \| \| `latency` \| 1722.228 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.41 pp sourcePopulation 81 stderr 1.405 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__emerging_companies_venture_capital/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [CorpFin v2 / overall](https://benchize.com/benchmarks/vals-ai%3Acorp_fin_v2%3Aoverall.md) | 71.83% | SE ±0.88 pp | 2 / 39 | $1.74 | 52.99source duration | [Vals AI · 2](https://www.vals.ai/_astro/benchmark_view_corp_fin_v2.CectKwl1.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 71.834 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 1.737573 \| USD \| per_test \| Lower is better \| \| `latency` \| 52.99 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.88 pp sourcePopulation 134 stderr 0.883 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_corp_fin_v2.CectKwl1.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 2 Observed 2026-10-21T16:45:57.998Z Evidence SHA-256 `7cf2217ab3728d5a84bffce82aab9fc52d7f810b0cbe8c65b387df47f3d31c9b` Source locator `/tasks/overall/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [Public Benefits Bench v1 / web search](https://benchize.com/benchmarks/vals-ai%3Apublic-benefits-bench-v1%3Aweb_search.md) | 67.73% | SE ±1.22 pp | 1 / 8 | $0.73 | 792.68source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_snap_v1.BPa0Hrre.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 67.727 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.729545 \| USD \| per_test \| Lower is better \| \| `latency` \| 792.684 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.22 pp sourcePopulation 13 stderr 1.216 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_snap_v1.BPa0Hrre.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:46:18.363Z Evidence SHA-256 `b85e2b3cbdf03e13e2ffdb27ce4f108003a548d705a24ae26a0a115e2027475d` Source locator `/tasks/web_search/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [Terminal-Bench 2.1 / easy](https://benchize.com/benchmarks/vals-ai%3Aterminal-bench-2-1%3Aeasy.md) | 100% | SE ±0 pp | 1 / 43 | $9.54 | 11,225.95source duration | [Vals AI · 2.1](https://www.vals.ai/_astro/benchmark_view_terminal_bench_2_1.CdyoObsA.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 100 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 9.538742 \| USD \| per_test \| Lower is better \| \| `latency` \| 11225.953 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0 pp sourcePopulation 76 stderr 0 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_terminal_bench_2_1.CdyoObsA.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 2.1 Observed 2026-10-21T16:46:13.216Z Evidence SHA-256 `a5306805cdf86495bdcceff6b2c8534a3768b9853f5077895dab975b809b054f` Source locator `/tasks/easy/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [APEX-Agents](https://benchize.com/benchmarks/apex-agents.md) | 43.3% | — | 1 / 6 | — | — | [Moonshot AI · sha256:0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c](https://www.kimi.ai/blog/kimi-k3) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 43.3 \| percent \| reported \| Higher is better \| ### Configuration Fallback policy unspecified fallbackEnabled true Reasoning effort max sourceFallbackLabel with fallback thinkingEnabled true Fallback is enabled for refusals; this does not mean every task used a fallback model. ### Source Score publisher [Moonshot AI](https://www.kimi.ai/blog/kimi-k3) Collected via [moonshot k3](https://www.kimi.ai/blog/kimi-k3) Publisher reference Referenced by the collection source; the upstream result has not been independently captured. Benchmark [https://www.kimi.ai/blog/kimi-k3](https://www.kimi.ai/blog/kimi-k3) [Collected via](https://www.kimi.ai/blog/kimi-k3) Source Moonshot AI Original model label Claude Fable 5 (max, with fallback) Source version sha256:0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c Observed 2026-10-21T16:31:44.130Z Evidence SHA-256 `0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c` Source locator `/flight/sheet/data/16/scores/1` |  |  |  |  |  |  |  |
| [Legal Research Bench / criminal](https://benchize.com/benchmarks/vals-ai%3Alegal_research%3Acriminal.md) | 42.11% | SE ±6.6 pp | 7 / 50 | $9.47 | 1,398.54source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_research.DiuL0R8P.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 42.105 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 9.468125 \| USD \| per_test \| Lower is better \| \| `latency` \| 1398.537 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±6.6 pp sourcePopulation 82 stderr 6.598 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_research.DiuL0R8P.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:52.389Z Evidence SHA-256 `26c58dfd47d0d2dc86f33c6d056f30309e0ee63570e7cd752c593b385f9d862e` Source locator `/tasks/criminal/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [Harvey's Legal Agent Benchmark v1 / criteria pass rate environmental esg](https://benchize.com/benchmarks/vals-ai%3Ahlab-v1%3Acriteria_pass_rate__environmental_esg.md) | 87.81% | SE ±1.96 pp | 8 / 49 | $17.04 | 1,679.81source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 87.814 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 17.039801 \| USD \| per_test \| Lower is better \| \| `latency` \| 1679.81 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.96 pp sourcePopulation 81 stderr 1.962 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__environmental_esg/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [Harvey's Legal Agent Benchmark v1 / criteria pass rate antitrust competition](https://benchize.com/benchmarks/vals-ai%3Ahlab-v1%3Acriteria_pass_rate__antitrust_competition.md) | 92.04% | SE ±1.81 pp | 6 / 49 | $16.83 | 1,501.21source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 92.035 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 16.832828 \| USD \| per_test \| Lower is better \| \| `latency` \| 1501.212 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.81 pp sourcePopulation 81 stderr 1.805 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__antitrust_competition/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [MMLU Pro / biology](https://benchize.com/benchmarks/vals-ai%3Ammlu_pro%3Abiology.md) | 93.58% | SE ±0.92 pp | 16 / 40 | $0.09 | 18.86source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_mmlu_pro.y416pXDi.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 93.584 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.086148 \| USD \| per_test \| Lower is better \| \| `latency` \| 18.862 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.92 pp sourcePopulation 138 stderr 0.915 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_mmlu_pro.y416pXDi.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:46:06.503Z Evidence SHA-256 `dab025507f46c716dd8d145e90ec0a737d9c56219853f41e4b1cc3c5e8c222b4` Source locator `/tasks/biology/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [FrontierMath Tier 4 v2 (private)](https://benchize.com/benchmarks/epoch-ai%3Afrontiermath_tier_4_v2.md) | 90.2% | SE ±4.6 pp | 7 / 38 | — | — | [epoch ai · FrontierMath-Tier-4-v2-Private](https://epoch.ai/data/benchmark_data.zip#file=frontiermath_tier_4_v2.csv) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 0.902 \| score_fraction \| Best score (across scorers) \| Higher is better \| \| `mean_score` \| 0.902 \| score_fraction \| mean_score \| Higher is better \| ### Configuration benchmarkVersion FrontierMath-Tier-4-v2-Private reasoningBudget Not reported Reasoning effort max sourceModelVersion claude-fable-5_max ### Reported statistics Uncertainty Standard error: ±4.6 pp standardError 0.046 ### Source [Original source](https://epoch.ai/data/benchmark_data.zip#file=frontiermath_tier_4_v2.csv) Source epoch ai Original model label claude-fable-5 Source version FrontierMath-Tier-4-v2-Private Evaluated 2026-06-09T18:12:35.000Z Observed 2026-10-21T16:30:24.199Z Evidence SHA-256 `48b886c828092231989fba6e3819efb7462e89ef681b7b306b4bfa3505829e4a` Source locator `/rows/75` |  |  |  |  |  |  |  |
| [WorldVQA ForceAnswer](https://benchize.com/benchmarks/worldvqa-force-answer.md) | 56.7% | — | 1 / 5 | — | — | [Moonshot AI · sha256:0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c](https://www.kimi.ai/blog/kimi-k3) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 56.7 \| percent \| reported \| Higher is better \| ### Configuration Fallback policy unspecified fallbackEnabled true forceAnswer true Reasoning effort max runs 3 sourceFallbackLabel with fallback thinkingEnabled true Fallback is enabled for refusals; this does not mean every task used a fallback model. ### Reported statistics runCount 3 ### Source Score publisher [Moonshot AI](https://www.kimi.ai/blog/kimi-k3) Collected via [moonshot k3](https://www.kimi.ai/blog/kimi-k3) Publisher reference Referenced by the collection source; the upstream result has not been independently captured. Benchmark [https://www.kimi.ai/blog/kimi-k3](https://www.kimi.ai/blog/kimi-k3) [Collected via](https://www.kimi.ai/blog/kimi-k3) Source Moonshot AI Original model label Claude Fable 5 (max, with fallback) Source version sha256:0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c Observed 2026-10-21T16:31:44.130Z Evidence SHA-256 `0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c` Source locator `/flight/sheet/data/32/scores/1` |  |  |  |  |  |  |  |
| [ArXivMath 05/2026](https://benchize.com/benchmarks/matharena%3Aarxiv--may.md) | 87.5% | 95% CI 81.58–93.42% | 5 / 21 | 3.91usd per problem | — | [matharena · arxiv--may](https://matharena.ai/competition_tables/arxiv--may) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 87.50 \| percent \| competition \| Higher is better \| \| `cost_per_problem` \| 3.91 \| usd_per_problem \| competition \| Lower is better \| ### Configuration competition arxiv--may predicted false Reasoning effort max sourceModelName Claude-Fable-5 (max) sourceModelVersion anthropic/fable_5_max thinkingEnabled true ### Reported statistics Uncertainty 95% confidence interval: 81.58–93.42% confidenceInterval95 5.92 problems 40 ### Source [Original source](https://matharena.ai/competition_tables/arxiv--may) Source matharena Original model label Claude-Fable-5 (max) Source version arxiv--may Observed 2026-10-21T16:52:29.369Z Evidence SHA-256 `758e1bba77e36f4ced2a51b4bc0d46bd956ed94ef6f57e93e1a974b75e76a4ce` Source locator `/table/rows/4` |  |  |  |  |  |  |  |
| [Tax Agent Bench / overall](https://benchize.com/benchmarks/vals-ai%3Atax_agent_bench%3Aoverall.md) | 38.28% | SE ±3.5 pp | 7 / 46 | $6.7 | 866.63source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_tax_agent_bench.BFpexnpS.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 38.281 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 6.696087 \| USD \| per_test \| Lower is better \| \| `latency` \| 866.628 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±3.5 pp sourcePopulation 73 stderr 3.503 tokenTotals {"cache_read_tokens":62150339,"cache_write_tokens":42027339,"input_tokens":104187949,"output_tokens":14106474,"reasoning_tokens":10318710} ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_tax_agent_bench.BFpexnpS.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:55.958Z Evidence SHA-256 `446db182b67efb7923b3187a4040ec3a3ea492b8ad699a2b6d708d714cc063ac` Source locator `/tasks/overall/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [Tax Agent Bench / weighted pass forms filings](https://benchize.com/benchmarks/vals-ai%3Atax_agent_bench%3Aweighted_pass__forms_filings.md) | 64% | SE ±8.45 pp | 14 / 46 | $9.61 | 1,114.89source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_tax_agent_bench.BFpexnpS.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 64.004 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 9.612732 \| USD \| per_test \| Lower is better \| \| `latency` \| 1114.894 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±8.45 pp sourcePopulation 73 stderr 8.454 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_tax_agent_bench.BFpexnpS.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:55.958Z Evidence SHA-256 `446db182b67efb7923b3187a4040ec3a3ea492b8ad699a2b6d708d714cc063ac` Source locator `/tasks/weighted_pass__forms_filings/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [SWE Marathon](https://benchize.com/benchmarks/swe-marathon.md) | 35% | — | 4 / 6 | — | — | [Moonshot AI · sha256:0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c](https://www.kimi.ai/blog/kimi-k3) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 35.0 \| percent \| reported \| Higher is better \| ### Configuration datasetAdaptation H20-calibrated Fallback policy unspecified fallbackEnabled true harness Claude Code Reasoning effort max sourceFallbackLabel with fallback thinkingEnabled true Fallback is enabled for refusals; this does not mean every task used a fallback model. ### Source Score publisher [Moonshot AI](https://www.kimi.ai/blog/kimi-k3) Collected via [moonshot k3](https://www.kimi.ai/blog/kimi-k3) Publisher reference Referenced by the collection source; the upstream result has not been independently captured. Benchmark [https://www.kimi.ai/blog/kimi-k3](https://www.kimi.ai/blog/kimi-k3) [Collected via](https://www.kimi.ai/blog/kimi-k3) Source Moonshot AI Original model label Claude Fable 5 (max, with fallback) Source version sha256:0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c Observed 2026-10-21T16:31:44.130Z Evidence SHA-256 `0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c` Source locator `/flight/sheet/data/4/scores/1` |  |  |  |  |  |  |  |
| [SAGE / automata](https://benchize.com/benchmarks/vals-ai%3Asage%3Aautomata.md) | 39.68% | SE ±6.18 pp | 12 / 41 | $0.78 | 152.56source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_sage.RRwIIzO5.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 39.682 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.779331 \| USD \| per_test \| Lower is better \| \| `latency` \| 152.558 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 30000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±6.18 pp sourcePopulation 91 stderr 6.183 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_sage.RRwIIzO5.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:46:07.515Z Evidence SHA-256 `cc5fb3fecdb0382d492ff167a42ad84b9c5c3883373c78ed87baaa972b8b6f7c` Source locator `/tasks/automata/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [Harvey's Legal Agent Benchmark v1 / white collar defense investigations](https://benchize.com/benchmarks/vals-ai%3Ahlab-v1%3Awhite_collar_defense_investigations.md) | 0% | SE ±0 pp | 1 / 49 | $16.63 | 1,555.02source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 0 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 16.626886 \| USD \| per_test \| Lower is better \| \| `latency` \| 1555.02 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0 pp sourcePopulation 81 stderr 0 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/white_collar_defense_investigations/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [Harvey's Legal Agent Benchmark v1 / international trade sanctions](https://benchize.com/benchmarks/vals-ai%3Ahlab-v1%3Ainternational_trade_sanctions.md) | 0% | SE ±0 pp | 4 / 49 | $15.32 | 1,640.22source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 0 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 15.317586 \| USD \| per_test \| Lower is better \| \| `latency` \| 1640.224 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0 pp sourcePopulation 81 stderr 0 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/international_trade_sanctions/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [Terminal-Bench 4.0 / ml](https://benchize.com/benchmarks/vals-ai%3Aterminal-bench-4%3Aml.md) | 42.42% | SE ±3.03 pp | 11 / 36 | — | — | [Vals AI · 4.0](https://www.vals.ai/_astro/benchmark_view_terminal_bench_4.CAnGyzC_.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 42.424 \| percent \| reported \| Higher is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±3.03 pp sourcePopulation 52 stderr 3.03 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_terminal_bench_4.CAnGyzC_.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 4.0 Observed 2026-10-21T16:46:10.642Z Evidence SHA-256 `98f6964580e901f1aa143db664bd92aa558415f643569492545d67c47d7f519b` Source locator `/tasks/ml/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [Vibe Code Bench v1.1 / overall](https://benchize.com/benchmarks/vals-ai%3Avibe-code%3Aoverall.md) | 90.35% | SE ±2.1 pp | 4 / 59 | $41.71 | 3,711.03source duration | [Vals AI · 1.1](https://www.vals.ai/_astro/benchmark_view_vibe_code_bench.DcEBMbCA.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 90.352 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 41.714694 \| USD \| per_test \| Lower is better \| \| `latency` \| 3711.026 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness OpenHands maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±2.1 pp sourcePopulation 116 stderr 2.097 tokenTotals {"cache_read_tokens":1604184760,"cache_write_tokens":24245069,"input_tokens":1628443000,"output_tokens":8412174,"reasoning_tokens":2231064} ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_vibe_code_bench.DcEBMbCA.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1.1 Observed 2026-10-21T16:46:08.625Z Evidence SHA-256 `0e8c23f521a0f751c5592952620e197b376cade41c78578461b7aa5740a173e7` Source locator `/tasks/overall/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [LiveCodeBench / medium](https://benchize.com/benchmarks/vals-ai%3Alcb%3Amedium.md) | 90.86% | SE ±1.47 pp | 4 / 43 | $0.19 | 61.05source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_lcb.miyRUvVa.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 90.862 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.194043 \| USD \| per_test \| Lower is better \| \| `latency` \| 61.049 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.47 pp sourcePopulation 143 stderr 1.472 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_lcb.miyRUvVa.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:46:14.220Z Evidence SHA-256 `1a7c7dabc9d95a0182886b1edc42623987cac2095a0d407a5e334f37acdfa2f5` Source locator `/tasks/medium/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [Harvey's Legal Agent Benchmark v1 / immigration](https://benchize.com/benchmarks/vals-ai%3Ahlab-v1%3Aimmigration.md) | 0% | SE ±0 pp | 6 / 49 | $14.2 | 1,489.5source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 0 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 14.198276 \| USD \| per_test \| Lower is better \| \| `latency` \| 1489.498 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0 pp sourcePopulation 81 stderr 0 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/immigration/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [DECK-Bench (Internal)](https://benchize.com/benchmarks/deck-bench.md) | 73% | — | 3 / 6 | — | — | [Moonshot AI · sha256:0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c](https://www.kimi.ai/blog/kimi-k3) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 73.0 \| percent \| reported \| Higher is better \| ### Configuration benchmarkVisibility internal Fallback policy unspecified fallbackEnabled true Reasoning effort max sourceFallbackLabel with fallback thinkingEnabled true Fallback is enabled for refusals; this does not mean every task used a fallback model. ### Source Score publisher [Moonshot AI](https://www.kimi.ai/blog/kimi-k3) Collected via [moonshot k3](https://www.kimi.ai/blog/kimi-k3) Publisher reference Referenced by the collection source; the upstream result has not been independently captured. Benchmark [https://www.kimi.ai/blog/kimi-k3](https://www.kimi.ai/blog/kimi-k3) [Collected via](https://www.kimi.ai/blog/kimi-k3) Source Moonshot AI Original model label Claude Fable 5 (max, with fallback) Source version sha256:0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c Observed 2026-10-21T16:31:44.130Z Evidence SHA-256 `0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c` Source locator `/flight/sheet/data/19/scores/1` |  |  |  |  |  |  |  |
| [Legal Research Bench / weighted pass constitutional civil rights](https://benchize.com/benchmarks/vals-ai%3Alegal_research%3Aweighted_pass__constitutional_civil_rights.md) | 87.91% | SE ±2.55 pp | 8 / 50 | $6.9 | 977.36source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_research.DiuL0R8P.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 87.91 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 6.898245 \| USD \| per_test \| Lower is better \| \| `latency` \| 977.356 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±2.55 pp sourcePopulation 82 stderr 2.549 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_research.DiuL0R8P.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:52.389Z Evidence SHA-256 `26c58dfd47d0d2dc86f33c6d056f30309e0ee63570e7cd752c593b385f9d862e` Source locator `/tasks/weighted_pass__constitutional_civil_rights/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [LiveCodeBench / easy](https://benchize.com/benchmarks/vals-ai%3Alcb%3Aeasy.md) | 98.76% | SE ±0.62 pp | 14 / 43 | $0.03 | 12.69source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_lcb.miyRUvVa.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 98.758 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.033455 \| USD \| per_test \| Lower is better \| \| `latency` \| 12.692 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.62 pp sourcePopulation 143 stderr 0.617 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_lcb.miyRUvVa.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:46:14.220Z Evidence SHA-256 `1a7c7dabc9d95a0182886b1edc42623987cac2095a0d407a5e334f37acdfa2f5` Source locator `/tasks/easy/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [LegalBench / interpretation tasks](https://benchize.com/benchmarks/vals-ai%3Alegal_bench%3Ainterpretation_tasks.md) | 86.86% | SE ±0.16 pp | 3 / 44 | $0.03 | 8.12source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 86.863 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.028526 \| USD \| per_test \| Lower is better \| \| `latency` \| 8.119 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.16 pp sourcePopulation 151 stderr 0.163 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/interpretation_tasks/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [Harvey's Legal Agent Benchmark v1 / criteria pass rate trusts estates private client](https://benchize.com/benchmarks/vals-ai%3Ahlab-v1%3Acriteria_pass_rate__trusts_estates_private_client.md) | 91.99% | SE ±1.68 pp | 5 / 49 | $18.19 | 1,662.66source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 91.985 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 18.186942 \| USD \| per_test \| Lower is better \| \| `latency` \| 1662.664 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.68 pp sourcePopulation 81 stderr 1.681 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__trusts_estates_private_client/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [EMB / template](https://benchize.com/benchmarks/vals-ai%3Aemb%3Atemplate.md) | 76.99% | SE ±2.63 pp | 5 / 47 | $12.64 | 1,549.9source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_emb.BY1LXsAw.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 76.991 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 12.636539 \| USD \| per_test \| Lower is better \| \| `latency` \| 1549.904 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±2.63 pp sourcePopulation 76 stderr 2.631 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_emb.BY1LXsAw.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:57.495Z Evidence SHA-256 `d807aee620ea47152a1fca0257a8dd12a1be4512c76cc1744c1ae9e285180918` Source locator `/tasks/template/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [Vals Web Search Index / legal](https://benchize.com/benchmarks/vals-ai%3Aweb_search_backends%3Alegal.md) | 49.04% | SE ±3.47 pp | 1 / 3 | $3.43 | 679.08source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_web_search_backends.DAfqOqhu.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 49.03846153846153 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 3.4339210841346155 \| USD \| per_test \| Lower is better \| \| `latency` \| 679.0761298076923 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Native maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±3.47 pp sourcePopulation 8 stderr 3.4745975638705247 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_web_search_backends.DAfqOqhu.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:50.373Z Evidence SHA-256 `8a022ad53b5c39bc71f346f44f4a1635122ecdca08efd420a41ddb2fba3082bb` Source locator `/tasks/legal/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [Legal Research Bench / civil litigation](https://benchize.com/benchmarks/vals-ai%3Alegal_research%3Acivil_litigation.md) | 44.44% | SE ±9.75 pp | 10 / 50 | $10.12 | 1,489.89source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_research.DiuL0R8P.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 44.444 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 10.123039 \| USD \| per_test \| Lower is better \| \| `latency` \| 1489.886 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±9.75 pp sourcePopulation 82 stderr 9.745 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_research.DiuL0R8P.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:52.389Z Evidence SHA-256 `26c58dfd47d0d2dc86f33c6d056f30309e0ee63570e7cd752c593b385f9d862e` Source locator `/tasks/civil_litigation/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [Vals Multimodal Index / finance agent](https://benchize.com/benchmarks/vals-ai%3Avals_multimodal_index%3Afinance_agent.md) | 56.35% | SE ±0.83 pp | 5 / 20 | $8.06 | 537.75source duration | [Vals AI · 1.2](https://www.vals.ai/_astro/benchmark_view_vals_multimodal_index.virVpWMs.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 56.349 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 8.061563 \| USD \| per_test \| Lower is better \| \| `latency` \| 537.752 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.83 pp sourcePopulation 33 stderr 0.831 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_vals_multimodal_index.virVpWMs.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1.2 Observed 2026-10-21T16:45:51.881Z Evidence SHA-256 `077b22575f6e075b7b78157d000ec8f94d7ead409c7dcbcf04887630ed572a2d` Source locator `/tasks/finance_agent/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [Legal Research Bench / health](https://benchize.com/benchmarks/vals-ai%3Alegal_research%3Ahealth.md) | 66.67% | SE ±8.75 pp | 15 / 50 | $12.38 | 1,126.58source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_research.DiuL0R8P.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 66.667 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 12.38182 \| USD \| per_test \| Lower is better \| \| `latency` \| 1126.576 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±8.75 pp sourcePopulation 82 stderr 8.754 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_research.DiuL0R8P.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:52.389Z Evidence SHA-256 `26c58dfd47d0d2dc86f33c6d056f30309e0ee63570e7cd752c593b385f9d862e` Source locator `/tasks/health/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [Legal Research Bench / weighted pass administrative regulatory](https://benchize.com/benchmarks/vals-ai%3Alegal_research%3Aweighted_pass__administrative_regulatory.md) | 90.22% | SE ±2.23 pp | 8 / 50 | $9.38 | 1,179.38source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_research.DiuL0R8P.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 90.224 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 9.381311 \| USD \| per_test \| Lower is better \| \| `latency` \| 1179.378 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±2.23 pp sourcePopulation 82 stderr 2.232 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_research.DiuL0R8P.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:52.389Z Evidence SHA-256 `26c58dfd47d0d2dc86f33c6d056f30309e0ee63570e7cd752c593b385f9d862e` Source locator `/tasks/weighted_pass__administrative_regulatory/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [Code Migration / cli](https://benchize.com/benchmarks/vals-ai%3Acode-migration%3Acli.md) | 60.09% | SE ±2.84 pp | 6 / 47 | $139.1 | 7,275.19source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_code_migration.Dttzoqk3.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 60.087 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 139.102342 \| USD \| per_test \| Lower is better \| \| `latency` \| 7275.185 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±2.84 pp sourcePopulation 79 stderr 2.841 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_code_migration.Dttzoqk3.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:46:12.712Z Evidence SHA-256 `99194597f2c76c402195be47839409046e18bd30454616da6dd40b736a088fa4` Source locator `/tasks/cli/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [Finance Agent (v2) / comparables](https://benchize.com/benchmarks/vals-ai%3Afabv2%3Acomparables.md) | 43.99% | SE ±0.72 pp | 29 / 50 | $10.48 | 812.32source duration | [Vals AI · 2](https://www.vals.ai/_astro/benchmark_view_finance_agent_v2.DypCOnwd.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 43.986 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 10.480979 \| USD \| per_test \| Lower is better \| \| `latency` \| 812.322 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.72 pp sourcePopulation 80 stderr 0.722 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_finance_agent_v2.DypCOnwd.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 2 Observed 2026-10-21T16:45:57.035Z Evidence SHA-256 `f731b20a26c126725d125519f2b5841aa45254656dba859a9570d7ce7e5d8191` Source locator `/tasks/comparables/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [Harvey's Legal Agent Benchmark v1 / capital markets](https://benchize.com/benchmarks/vals-ai%3Ahlab-v1%3Acapital_markets.md) | 0% | SE ±0 pp | 16 / 49 | $17.51 | 1,359.53source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 0 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 17.509413 \| USD \| per_test \| Lower is better \| \| `latency` \| 1359.526 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0 pp sourcePopulation 81 stderr 0 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/capital_markets/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [LegalBench / conclusion tasks](https://benchize.com/benchmarks/vals-ai%3Alegal_bench%3Aconclusion_tasks.md) | 91.78% | SE ±0.67 pp | 2 / 44 | $0.02 | 7.26source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 91.775 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.019734 \| USD \| per_test \| Lower is better \| \| `latency` \| 7.255 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.67 pp sourcePopulation 151 stderr 0.667 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [CorpFin v2 / shared max context](https://benchize.com/benchmarks/vals-ai%3Acorp_fin_v2%3Ashared_max_context.md) | 71.56% | SE ±1.54 pp | 4 / 39 | $1.39 | 58.9source duration | [Vals AI · 2](https://www.vals.ai/_astro/benchmark_view_corp_fin_v2.CectKwl1.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 71.562 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 1.387598 \| USD \| per_test \| Lower is better \| \| `latency` \| 58.898 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.54 pp sourcePopulation 134 stderr 1.54 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_corp_fin_v2.CectKwl1.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 2 Observed 2026-10-21T16:45:57.998Z Evidence SHA-256 `7cf2217ab3728d5a84bffce82aab9fc52d7f810b0cbe8c65b387df47f3d31c9b` Source locator `/tasks/shared_max_context/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [MMLU Pro / computer science](https://benchize.com/benchmarks/vals-ai%3Ammlu_pro%3Acomputer_science.md) | 94.15% | SE ±1.16 pp | 2 / 40 | $0.08 | 17.47source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_mmlu_pro.y416pXDi.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 94.146 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.08492 \| USD \| per_test \| Lower is better \| \| `latency` \| 17.469 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.16 pp sourcePopulation 138 stderr 1.159 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_mmlu_pro.y416pXDi.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:46:06.503Z Evidence SHA-256 `dab025507f46c716dd8d145e90ec0a737d9c56219853f41e4b1cc3c5e8c222b4` Source locator `/tasks/computer_science/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [Harvey's Legal Agent Benchmark v1 / arbitration international dispute resolution](https://benchize.com/benchmarks/vals-ai%3Ahlab-v1%3Aarbitration_international_dispute_resolution.md) | 0% | SE ±0 pp | 4 / 49 | $22.49 | 1,715.52source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 0 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 22.492624 \| USD \| per_test \| Lower is better \| \| `latency` \| 1715.52 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0 pp sourcePopulation 81 stderr 0 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/arbitration_international_dispute_resolution/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [MMLU Pro / psychology](https://benchize.com/benchmarks/vals-ai%3Ammlu_pro%3Apsychology.md) | 92.23% | SE ±0.95 pp | 3 / 40 | $0.05 | 13.23source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_mmlu_pro.y416pXDi.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 92.231 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.04971 \| USD \| per_test \| Lower is better \| \| `latency` \| 13.228 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.95 pp sourcePopulation 138 stderr 0.948 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_mmlu_pro.y416pXDi.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:46:06.503Z Evidence SHA-256 `dab025507f46c716dd8d145e90ec0a737d9c56219853f41e4b1cc3c5e8c222b4` Source locator `/tasks/psychology/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [Legal Research Bench / administrative regulatory](https://benchize.com/benchmarks/vals-ai%3Alegal_research%3Aadministrative_regulatory.md) | 57.58% | SE ±6.13 pp | 10 / 50 | $9.38 | 1,179.38source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_research.DiuL0R8P.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 57.576 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 9.381311 \| USD \| per_test \| Lower is better \| \| `latency` \| 1179.378 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±6.13 pp sourcePopulation 82 stderr 6.13 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_research.DiuL0R8P.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:52.389Z Evidence SHA-256 `26c58dfd47d0d2dc86f33c6d056f30309e0ee63570e7cd752c593b385f9d862e` Source locator `/tasks/administrative_regulatory/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [SWE-bench / <15 min fix](https://benchize.com/benchmarks/vals-ai%3Aswebench%3A%3C15%20min%20fix.md) | 95.88% | SE ±1.43 pp | 6 / 39 | $0.41 | 260.53source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_swebench.Ae8ni7h0.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 95.876 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.410248 \| USD \| per_test \| Lower is better \| \| `latency` \| 260.529 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.43 pp sourcePopulation 88 stderr 1.431 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_swebench.Ae8ni7h0.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:46:13.711Z Evidence SHA-256 `7a4e2e8485283caac500127419f68ec4bd5414c837ade114fc6afa5b2d34ccaa` Source locator `/tasks/<15 min fix/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [HLE-Full w/ tools](https://benchize.com/benchmarks/hle-full-tools.md) | 63% | — | 1 / 5 | — | — | [Moonshot AI · sha256:0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c](https://www.kimi.ai/blog/kimi-k3) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 63.0 \| percent \| reported \| Higher is better \| ### Configuration Fallback policy unspecified fallbackEnabled true Reasoning effort max sourceFallbackLabel with fallback thinkingEnabled true toolsEnabled true Fallback is enabled for refusals; this does not mean every task used a fallback model. ### Source Score publisher [Moonshot AI](https://www.kimi.ai/blog/kimi-k3) Collected via [moonshot k3](https://www.kimi.ai/blog/kimi-k3) Publisher reference Referenced by the collection source; the upstream result has not been independently captured. Benchmark [https://www.kimi.ai/blog/kimi-k3](https://www.kimi.ai/blog/kimi-k3) [Collected via](https://www.kimi.ai/blog/kimi-k3) Source Moonshot AI Original model label Claude Fable 5 (max, with fallback) Source version sha256:0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c Observed 2026-10-21T16:31:44.130Z Evidence SHA-256 `0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c` Source locator `/flight/sheet/data/22/scores/1` |  |  |  |  |  |  |  |
| [Terminal-Bench 4.0 / science](https://benchize.com/benchmarks/vals-ai%3Aterminal-bench-4%3Ascience.md) | 45.24% | SE ±2.38 pp | 9 / 36 | — | — | [Vals AI · 4.0](https://www.vals.ai/_astro/benchmark_view_terminal_bench_4.CAnGyzC_.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 45.238 \| percent \| reported \| Higher is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±2.38 pp sourcePopulation 52 stderr 2.381 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_terminal_bench_4.CAnGyzC_.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 4.0 Observed 2026-10-21T16:46:10.642Z Evidence SHA-256 `98f6964580e901f1aa143db664bd92aa558415f643569492545d67c47d7f519b` Source locator `/tasks/science/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [MortgageTax / overall](https://benchize.com/benchmarks/vals-ai%3Amortgage_tax%3Aoverall.md) | 68.92% | SE ±0.91 pp | 7 / 31 | $0.18 | 16.1source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_mortgage_tax.DCd-oO5Y.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 68.919 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.175013 \| USD \| per_test \| Lower is better \| \| `latency` \| 16.096 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.91 pp sourcePopulation 98 stderr 0.909 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_mortgage_tax.DCd-oO5Y.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:58.500Z Evidence SHA-256 `e5c844561f97d119d66d70269e726ecfb68be79eae39acf14ead3e95422e648c` Source locator `/tasks/overall/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [Job Bench](https://benchize.com/benchmarks/jobbench.md) | 57.4% | — | 1 / 6 | — | — | [Moonshot AI · sha256:0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c](https://www.kimi.ai/blog/kimi-k3) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 57.4 \| percent \| reported \| Higher is better \| ### Configuration Fallback policy unspecified fallbackEnabled true Reasoning effort max sourceFallbackLabel with fallback thinkingEnabled true Fallback is enabled for refusals; this does not mean every task used a fallback model. ### Source Score publisher [Moonshot AI](https://www.kimi.ai/blog/kimi-k3) Collected via [moonshot k3](https://www.kimi.ai/blog/kimi-k3) Publisher reference Referenced by the collection source; the upstream result has not been independently captured. Benchmark [https://www.kimi.ai/blog/kimi-k3](https://www.kimi.ai/blog/kimi-k3) [Collected via](https://www.kimi.ai/blog/kimi-k3) Source Moonshot AI Original model label Claude Fable 5 (max, with fallback) Source version sha256:0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c Observed 2026-10-21T16:31:44.130Z Evidence SHA-256 `0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c` Source locator `/flight/sheet/data/14/scores/1` |  |  |  |  |  |  |  |
| [Harvey's Legal Agent Benchmark v1 / criteria pass rate white collar defense investigations](https://benchize.com/benchmarks/vals-ai%3Ahlab-v1%3Acriteria_pass_rate__white_collar_defense_investigations.md) | 87.21% | SE ±1.94 pp | 12 / 49 | $16.63 | 1,555.02source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 87.205 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 16.626886 \| USD \| per_test \| Lower is better \| \| `latency` \| 1555.02 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.94 pp sourcePopulation 81 stderr 1.942 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__white_collar_defense_investigations/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [Kimi Code Bench 2.0 (Internal)](https://benchize.com/benchmarks/kimi-code-bench-2.md) | 76.9% | — | 1 / 6 | — | — | [Moonshot AI · sha256:0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c](https://www.kimi.ai/blog/kimi-k3) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 76.9 \| percent \| reported \| Higher is better \| ### Configuration benchmarkVisibility internal Fallback policy unspecified fallbackEnabled true harness Claude Code Reasoning effort max sourceFallbackLabel with fallback thinkingEnabled true Fallback is enabled for refusals; this does not mean every task used a fallback model. ### Source Score publisher [Moonshot AI](https://www.kimi.ai/blog/kimi-k3) Collected via [moonshot k3](https://www.kimi.ai/blog/kimi-k3) Publisher reference Referenced by the collection source; the upstream result has not been independently captured. Benchmark [https://www.kimi.ai/blog/kimi-k3](https://www.kimi.ai/blog/kimi-k3) [Collected via](https://www.kimi.ai/blog/kimi-k3) Source Moonshot AI Original model label Claude Fable 5 (max, with fallback) Source version sha256:0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c Observed 2026-10-21T16:31:44.130Z Evidence SHA-256 `0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c` Source locator `/flight/sheet/data/7/scores/1` |  |  |  |  |  |  |  |
| [Public Benefits Bench v1 / multi turn](https://benchize.com/benchmarks/vals-ai%3Apublic-benefits-bench-v1%3Amulti_turn.md) | 65.22% | SE ±1.24 pp | 1 / 8 | $0.29 | 272.42source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_snap_v1.BPa0Hrre.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 65.223 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.294256 \| USD \| per_test \| Lower is better \| \| `latency` \| 272.419 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.24 pp sourcePopulation 13 stderr 1.239 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_snap_v1.BPa0Hrre.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:46:18.363Z Evidence SHA-256 `b85e2b3cbdf03e13e2ffdb27ce4f108003a548d705a24ae26a0a115e2027475d` Source locator `/tasks/multi_turn/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [Finance Agent (v2) / overall](https://benchize.com/benchmarks/vals-ai%3Afabv2%3Aoverall.md) | 56.31% | SE ±0.84 pp | 16 / 50 | $8.06 | 610.95source duration | [Vals AI · 2](https://www.vals.ai/_astro/benchmark_view_finance_agent_v2.DypCOnwd.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 56.314 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 8.061563 \| USD \| per_test \| Lower is better \| \| `latency` \| 610.952 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.84 pp sourcePopulation 80 stderr 0.84 tokenTotals {"cache_read_tokens":110357130,"cache_write_tokens":201446408,"input_tokens":311882539,"output_tokens":19969523,"reasoning_tokens":null} ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_finance_agent_v2.DypCOnwd.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 2 Observed 2026-10-21T16:45:57.035Z Evidence SHA-256 `f731b20a26c126725d125519f2b5841aa45254656dba859a9570d7ce7e5d8191` Source locator `/tasks/overall/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [Vals Web Search Index / finance](https://benchize.com/benchmarks/vals-ai%3Aweb_search_backends%3Afinance.md) | 44.84% | SE ±2.36 pp | 3 / 3 | $4.64 | 805.82source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_web_search_backends.DAfqOqhu.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 44.83765024036157 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 4.635607564676617 \| USD \| per_test \| Lower is better \| \| `latency` \| 805.821472636816 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Native maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±2.36 pp sourcePopulation 8 stderr 2.3588410103167337 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_web_search_backends.DAfqOqhu.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:50.373Z Evidence SHA-256 `8a022ad53b5c39bc71f346f44f4a1635122ecdca08efd420a41ddb2fba3082bb` Source locator `/tasks/finance/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [HLE-Full](https://benchize.com/benchmarks/hle-full.md) | 53.3% | — | 1 / 5 | — | — | [Moonshot AI · sha256:0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c](https://www.kimi.ai/blog/kimi-k3) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 53.3 \| percent \| reported \| Higher is better \| ### Configuration Fallback policy unspecified fallbackEnabled true Reasoning effort max sourceFallbackLabel with fallback thinkingEnabled true toolsEnabled false Fallback is enabled for refusals; this does not mean every task used a fallback model. ### Source Score publisher [Moonshot AI](https://www.kimi.ai/blog/kimi-k3) Collected via [moonshot k3](https://www.kimi.ai/blog/kimi-k3) Publisher reference Referenced by the collection source; the upstream result has not been independently captured. Benchmark [https://www.kimi.ai/blog/kimi-k3](https://www.kimi.ai/blog/kimi-k3) [Collected via](https://www.kimi.ai/blog/kimi-k3) Source Moonshot AI Original model label Claude Fable 5 (max, with fallback) Source version sha256:0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c Observed 2026-10-21T16:31:44.130Z Evidence SHA-256 `0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c` Source locator `/flight/sheet/data/21/scores/1` |  |  |  |  |  |  |  |
| [Terminal-Bench 4.0 / security](https://benchize.com/benchmarks/vals-ai%3Aterminal-bench-4%3Asecurity.md) | 60% | SE ±0 pp | 10 / 36 | — | — | [Vals AI · 4.0](https://www.vals.ai/_astro/benchmark_view_terminal_bench_4.CAnGyzC_.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 60 \| percent \| reported \| Higher is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0 pp sourcePopulation 52 stderr 0 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_terminal_bench_4.CAnGyzC_.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 4.0 Observed 2026-10-21T16:46:10.642Z Evidence SHA-256 `98f6964580e901f1aa143db664bd92aa558415f643569492545d67c47d7f519b` Source locator `/tasks/security/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [Harvey's Legal Agent Benchmark v1 / insurance](https://benchize.com/benchmarks/vals-ai%3Ahlab-v1%3Ainsurance.md) | 0% | SE ±0 pp | 2 / 49 | $15.61 | 1,417.02source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 0 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 15.611186 \| USD \| per_test \| Lower is better \| \| `latency` \| 1417.02 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0 pp sourcePopulation 81 stderr 0 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/insurance/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [Finance Agent (v2) / earnings analysis](https://benchize.com/benchmarks/vals-ai%3Afabv2%3Aearnings_analysis.md) | 77.26% | SE ±1.28 pp | 9 / 50 | $2.99 | 369.25source duration | [Vals AI · 2](https://www.vals.ai/_astro/benchmark_view_finance_agent_v2.DypCOnwd.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 77.263 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 2.990129 \| USD \| per_test \| Lower is better \| \| `latency` \| 369.254 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.28 pp sourcePopulation 80 stderr 1.276 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_finance_agent_v2.DypCOnwd.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 2 Observed 2026-10-21T16:45:57.035Z Evidence SHA-256 `f731b20a26c126725d125519f2b5841aa45254656dba859a9570d7ce7e5d8191` Source locator `/tasks/earnings_analysis/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [GPQA Diamond](https://benchize.com/benchmarks/gpqa-diamond.md) | 92.6% | — | 4 / 6 | — | — | [Moonshot AI · sha256:0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c](https://www.kimi.ai/blog/kimi-k3) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 92.6 \| percent \| reported \| Higher is better \| ### Configuration Fallback policy unspecified fallbackEnabled true Reasoning effort max sourceFallbackLabel with fallback thinkingEnabled true Fallback is enabled for refusals; this does not mean every task used a fallback model. ### Source Score publisher [Moonshot AI](https://www.kimi.ai/blog/kimi-k3) Collected via [moonshot k3](https://www.kimi.ai/blog/kimi-k3) Publisher reference Referenced by the collection source; the upstream result has not been independently captured. Benchmark [https://www.kimi.ai/blog/kimi-k3](https://www.kimi.ai/blog/kimi-k3) [Collected via](https://www.kimi.ai/blog/kimi-k3) Source Moonshot AI Original model label Claude Fable 5 (max, with fallback) Source version sha256:0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c Observed 2026-10-21T16:31:44.130Z Evidence SHA-256 `0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c` Source locator `/flight/sheet/data/20/scores/1` |  |  |  |  |  |  |  |
| [Terminal-Bench 2.1 / medium](https://benchize.com/benchmarks/vals-ai%3Aterminal-bench-2-1%3Amedium.md) | 84.85% | SE ±0.61 pp | 6 / 43 | $0.69 | 816.43source duration | [Vals AI · 2.1](https://www.vals.ai/_astro/benchmark_view_terminal_bench_2_1.CdyoObsA.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 84.848 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.693727 \| USD \| per_test \| Lower is better \| \| `latency` \| 816.433 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.61 pp sourcePopulation 76 stderr 0.606 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_terminal_bench_2_1.CdyoObsA.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 2.1 Observed 2026-10-21T16:46:13.216Z Evidence SHA-256 `a5306805cdf86495bdcceff6b2c8534a3768b9853f5077895dab975b809b054f` Source locator `/tasks/medium/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [Harvey's Legal Agent Benchmark v1 / criteria pass rate litigation dispute resolution](https://benchize.com/benchmarks/vals-ai%3Ahlab-v1%3Acriteria_pass_rate__litigation_dispute_resolution.md) | 89.14% | SE ±1.91 pp | 11 / 49 | $16.21 | 1,467.62source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 89.139 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 16.209915 \| USD \| per_test \| Lower is better \| \| `latency` \| 1467.622 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.91 pp sourcePopulation 81 stderr 1.908 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__litigation_dispute_resolution/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [BrowseComp](https://benchize.com/benchmarks/browsecomp.md) | 88% | — | 3 / 5 | — | — | [Moonshot AI · sha256:0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c](https://www.kimi.ai/blog/kimi-k3) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 88.0 \| percent \| reported \| Higher is better \| ### Configuration Fallback policy unspecified fallbackEnabled true Reasoning effort max sourceFallbackLabel with fallback thinkingEnabled true Fallback is enabled for refusals; this does not mean every task used a fallback model. ### Source Score publisher [Moonshot AI](https://www.kimi.ai/blog/kimi-k3) Collected via [moonshot k3](https://www.kimi.ai/blog/kimi-k3) Publisher reference Referenced by the collection source; the upstream result has not been independently captured. Benchmark [https://www.kimi.ai/blog/kimi-k3](https://www.kimi.ai/blog/kimi-k3) [Collected via](https://www.kimi.ai/blog/kimi-k3) Source Moonshot AI Original model label Claude Fable 5 (max, with fallback) Source version sha256:0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c Observed 2026-10-21T16:31:44.130Z Evidence SHA-256 `0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c` Source locator `/flight/sheet/data/9/scores/1` |  |  |  |  |  |  |  |
| [Harvey's Legal Agent Benchmark v1 / criteria pass rate corporate governance](https://benchize.com/benchmarks/vals-ai%3Ahlab-v1%3Acriteria_pass_rate__corporate_governance.md) | 92.12% | SE ±1.58 pp | 9 / 49 | $18.66 | 1,644.85source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 92.123 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 18.661957 \| USD \| per_test \| Lower is better \| \| `latency` \| 1644.848 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.58 pp sourcePopulation 81 stderr 1.579 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__corporate_governance/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [Harvey's Legal Agent Benchmark v1 / funds asset management](https://benchize.com/benchmarks/vals-ai%3Ahlab-v1%3Afunds_asset_management.md) | 0% | SE ±0 pp | 11 / 49 | $32.83 | 2,223.2source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 0 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 32.834156 \| USD \| per_test \| Lower is better \| \| `latency` \| 2223.198 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0 pp sourcePopulation 81 stderr 0 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/funds_asset_management/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [Terminal-Bench 4.0 / operations](https://benchize.com/benchmarks/vals-ai%3Aterminal-bench-4%3Aoperations.md) | 25.93% | SE ±9.8 pp | 10 / 36 | — | — | [Vals AI · 4.0](https://www.vals.ai/_astro/benchmark_view_terminal_bench_4.CAnGyzC_.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 25.926 \| percent \| reported \| Higher is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±9.8 pp sourcePopulation 52 stderr 9.799 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_terminal_bench_4.CAnGyzC_.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 4.0 Observed 2026-10-21T16:46:10.642Z Evidence SHA-256 `98f6964580e901f1aa143db664bd92aa558415f643569492545d67c47d7f519b` Source locator `/tasks/operations/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [Harvey's Legal Agent Benchmark v1 / real estate](https://benchize.com/benchmarks/vals-ai%3Ahlab-v1%3Areal_estate.md) | 0% | SE ±0 pp | 6 / 49 | $21.07 | 1,678.2source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 0 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 21.067576 \| USD \| per_test \| Lower is better \| \| `latency` \| 1678.198 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0 pp sourcePopulation 81 stderr 0 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/real_estate/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [BabyVision w/ python](https://benchize.com/benchmarks/babyvision-python.md) | 90.5% | — | 1 / 5 | — | — | [Moonshot AI · sha256:0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c](https://www.kimi.ai/blog/kimi-k3) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 90.5 \| percent \| reported \| Higher is better \| ### Configuration Fallback policy unspecified fallbackEnabled true Reasoning effort max runs 3 sourceFallbackLabel with fallback thinkingEnabled true tools ["python"] toolsEnabled true Fallback is enabled for refusals; this does not mean every task used a fallback model. ### Reported statistics runCount 3 ### Source Score publisher [Moonshot AI](https://www.kimi.ai/blog/kimi-k3) Collected via [moonshot k3](https://www.kimi.ai/blog/kimi-k3) Publisher reference Referenced by the collection source; the upstream result has not been independently captured. Benchmark [https://www.kimi.ai/blog/kimi-k3](https://www.kimi.ai/blog/kimi-k3) [Collected via](https://www.kimi.ai/blog/kimi-k3) Source Moonshot AI Original model label Claude Fable 5 (max, with fallback) Source version sha256:0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c Observed 2026-10-21T16:31:44.130Z Evidence SHA-256 `0e806bbad76dc8e8dbc899a86051e3406bdd5995fe6c6aa1b2176dde4fb28e6c` Source locator `/flight/sheet/data/29/scores/1` |  |  |  |  |  |  |  |
| [Harvey's Legal Agent Benchmark v1 / criteria pass rate employment labor](https://benchize.com/benchmarks/vals-ai%3Ahlab-v1%3Acriteria_pass_rate__employment_labor.md) | 87.64% | SE ±2.02 pp | 5 / 49 | $14.37 | 1,459.55source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 87.64 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 14.366748 \| USD \| per_test \| Lower is better \| \| `latency` \| 1459.546 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±2.02 pp sourcePopulation 81 stderr 2.018 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_agent_benchmark_v1.ZWRzZKcv.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__employment_labor/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [FrontierCode 1.1 main](https://benchize.com/benchmarks/frontiercode%3Av1_1%3Amain.md) | 51.57% | — | 9 / 125 | $19.07 | — | [frontiercode](https://cognition.com/data/frontiercode-leaderboard/data.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `score` \| 0.5157 \| score_fraction \| blocking_criteria_adjusted \| Higher is better \| \| `correct` \| 0.5689 \| score_fraction \| raw_rubric \| Higher is better \| \| `cost` \| 19.0696 \| USD \| per_rollout \| Lower is better \| ### Configuration harness claude-code Reasoning effort max sourceVersion v1_1 subset main ### Reported statistics flaggedRate 0.002 sampleSize 100 ### Source [Original source](https://cognition.com/data/frontiercode-leaderboard/data.json) Source frontiercode Original model label Claude Fable 5 Observed 2026-10-21T16:31:00.878Z Evidence SHA-256 `ff0040778b7bdf69c705ad7474f771fb5e53cfaf5cf8573b3a4614153543d75f` Source locator `/v1_1/data/Claude%20Fable%205/max/main` |  |  |  |  |  |  |  |
| [Terminal-Bench 4.0 / media](https://benchize.com/benchmarks/vals-ai%3Aterminal-bench-4%3Amedia.md) | 50% | SE ±0 pp | 7 / 36 | — | — | [Vals AI · 4.0](https://www.vals.ai/_astro/benchmark_view_terminal_bench_4.CAnGyzC_.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 50 \| percent \| reported \| Higher is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0 pp sourcePopulation 52 stderr 0 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_terminal_bench_4.CAnGyzC_.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 4.0 Observed 2026-10-21T16:46:10.642Z Evidence SHA-256 `98f6964580e901f1aa143db664bd92aa558415f643569492545d67c47d7f519b` Source locator `/tasks/media/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |
| [APEX-Agents](https://benchize.com/benchmarks/mercor%3Aapex-agents%3A1.1.md) | 63.6% | — | 9 / 38 | — | — | [mercor · 1.1](https://www.mercor.com/apex/apex-agents-leaderboard/) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `pass_1` \| 63.6 \| percent \| task_completion_pass_at_1 \| Higher is better \| \| `mean_score` \| 74.5 \| percent \| rubric_mean \| Higher is better \| ### Configuration datasetUrl https://huggingface.co/datasets/mercor/apex-agents-v1.1 datasetVersion 1.1 harness loop_truncated_tools_agent Reasoning effort max ### Reported statistics reportedErrors {"mean_score":4.3,"pass_1":5.3} sampleSize 959 ### Source [Original source](https://www.mercor.com/apex/apex-agents-leaderboard/) Source mercor Original model label Fable 5 Source version 1.1 Observed 2026-10-21T16:33:27.509Z Evidence SHA-256 `359c569c3b0b62ab375895eabf3c45846fa45551d419d5c9d6a4c9117db78184` Source locator `/__NEXT_DATA__/props/pageProps/leaderboardData/6` |  |  |  |  |  |  |  |
| [MMLU Pro / engineering](https://benchize.com/benchmarks/vals-ai%3Ammlu_pro%3Aengineering.md) | 89.99% | SE ±0.96 pp | 2 / 40 | $0.19 | 80.54source duration | [Vals AI · 1](https://www.vals.ai/_astro/benchmark_view_mmlu_pro.y416pXDi.json) |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 89.99 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.189415 \| USD \| per_test \| Lower is better \| \| `latency` \| 80.54 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.96 pp sourcePopulation 138 stderr 0.964 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_mmlu_pro.y416pXDi.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:46:06.503Z Evidence SHA-256 `dab025507f46c716dd8d145e90ec0a737d9c56219853f41e4b1cc3c5e8c222b4` Source locator `/tasks/engineering/anthropic~1claude-fable-5` |  |  |  |  |  |  |  |

[Next page](https://benchize.com/models/claude-fable-5-max.md?cursor=6f3fdf358c5611c1eb959e1b737cf6d91f62e14b239286fc0ac8a41983444f58)

---
Cite as: Claude Fable 5Max — Benchize, https://benchize.com/models/claude-fable-5-max (retrieved 2026-10-21)