# LegalBench / conclusion tasks > Source: Benchize (https://benchize.com) — https://benchize.com/benchmarks/vals-ai%3Alegal_bench%3Aconclusion_tasks. When citing or reusing this content, credit "Benchize" and link to https://benchize.com/benchmarks/vals-ai%3Alegal_bench%3Aconclusion_tasks. Evaluating language models on a wide range of open source legal reasoning tasks. Canonical URL: https://benchize.com/benchmarks/vals-ai%3Alegal_bench%3Aconclusion_tasks Last updated: Wed, 21 Oct 2026 18:52:20 GMT Reproduction and quotation must name Benchize, link to the canonical HTML URL, and retain individual source citations. Evaluating language models on a wide range of open source legal reasoning tasks. Professional work · Models tested: 44 · Score publishers: Vals AI ## Model results | # | Model | Score | ± | Cost | Latency | | | --- | --- | --- | --- | --- | --- | --- | | 1 | [Claude Fable 5.1](https://benchize.com/models/claude-fable-5-1-max.md)Max | 92.51% | SE ±0.79 pp | $0.02 | 6.9source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 92.511 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.023415 \| USD \| per_test \| Lower is better \| \| `latency` \| 6.903 \| 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: ±0.79 pp sourcePopulation 151 stderr 0.79 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-5-1 Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/anthropic~1claude-fable-5-1` | | | | | | | | 2 | [Claude Fable 5](https://benchize.com/models/claude-fable-5-max.md)Max | 91.78% | SE ±0.67 pp | $0.02 | 7.26source duration | | | ### 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` | | | | | | | | 3 | [Claude Opus 5](https://benchize.com/models/claude-opus-5-max.md)Max | 91.56% | SE ±0.7 pp | $0.00538 | 2.37source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 91.56 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.005379 \| USD \| per_test \| Lower is better \| \| `latency` \| 2.373 \| 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: ±0.7 pp sourcePopulation 151 stderr 0.696 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-opus-5 Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/anthropic~1claude-opus-5` | | | | | | | | 4 | [Gemini 3.7 Flash](https://benchize.com/models/gemini-3-7-flash-high.md)High | 91.15% | SE ±0.89 pp | $0.00287 | 1.88source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 91.146 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.002867 \| USD \| per_test \| Lower is better \| \| `latency` \| 1.884 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.89 pp sourcePopulation 151 stderr 0.891 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label google/gemini-3.7-flash Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/google~1gemini-3.7-flash` | | | | | | | | 5 | [Gemini 3.6 Flash](https://benchize.com/models/gemini-3-6-flash-high.md)High | 91.08% | SE ±0.66 pp | $0.00395 | 2.69source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 91.081 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.003949 \| USD \| per_test \| Lower is better \| \| `latency` \| 2.694 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.66 pp sourcePopulation 151 stderr 0.662 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label google/gemini-3.6-flash Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/google~1gemini-3.6-flash` | | | | | | | | 6 | [Gemini 4 Argon](https://benchize.com/models/gemini-4-argon-high.md)High | 90.85% | SE ±0.61 pp | $0.01 | 9.58source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 90.847 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.011481 \| USD \| per_test \| Lower is better \| \| `latency` \| 9.579 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 262144 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.61 pp sourcePopulation 151 stderr 0.613 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label google/gemini-4-argon Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/google~1gemini-4-argon` | | | | | | | | 7 | [Gemini 3.8 Flash](https://benchize.com/models/gemini-3-8-flash-high.md)High | 90.69% | SE ±0.95 pp | $0.00523 | 2.43source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 90.691 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.005226 \| USD \| per_test \| Lower is better \| \| `latency` \| 2.432 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.95 pp sourcePopulation 151 stderr 0.949 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/google~1gemini-3.8-flash` | | | | | | | | 8 | [GPT-5.6 Sol](https://benchize.com/models/gpt-5-6-sol-max.md)Max | 90.49% | SE ±0.94 pp | $0.00562 | 2.88source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 90.492 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.005618 \| USD \| per_test \| Lower is better \| \| `latency` \| 2.88 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 128000 provider OpenAI reasoning Not reported Reasoning effort max Temperature Not reported topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.94 pp sourcePopulation 151 stderr 0.937 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label openai/gpt-5.6-sol Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/openai~1gpt-5.6-sol` | | | | | | | | 9 | [GPT-5.6 Terra](https://benchize.com/models/gpt-5-6-terra-max.md)Max | 90.33% | SE ±0.92 pp | $0.00185 | 2.37source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 90.327 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.001848 \| USD \| per_test \| Lower is better \| \| `latency` \| 2.373 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 128000 provider OpenAI reasoning Not reported Reasoning effort max Temperature Not reported topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.92 pp sourcePopulation 151 stderr 0.916 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label openai/gpt-5.6-terra Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/openai~1gpt-5.6-terra` | | | | | | | | 10 | [Gemini 3 Flash Preview](https://benchize.com/models/gemini-3-flash-preview-high.md)High | 90.08% | SE ±0.61 pp | $0.00174 | 4.23source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 90.077 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.001736 \| USD \| per_test \| Lower is better \| \| `latency` \| 4.23 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.61 pp sourcePopulation 151 stderr 0.613 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label google/gemini-3-flash-preview Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/google~1gemini-3-flash-preview` | | | | | | | | 11 | [Claude Opus 4.7](https://benchize.com/models/claude-opus-4-7-max.md)Max | 89.98% | SE ±0.93 pp | $0.00707 | 16.72source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 89.983 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.007066 \| USD \| per_test \| Lower is better \| \| `latency` \| 16.717 \| 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 151 stderr 0.934 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-opus-4-7 Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/anthropic~1claude-opus-4-7` | | | | | | | | 12 | [GPT-5.5](https://benchize.com/models/gpt-5-5-xhigh.md)Extra high | 89.64% | SE ±0.7 pp | $0.00784 | 11.17source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 89.644 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.00784 \| USD \| per_test \| Lower is better \| \| `latency` \| 11.168 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 128000 provider OpenAI reasoning Not reported Reasoning effort xhigh Temperature Not reported topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.7 pp sourcePopulation 151 stderr 0.697 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label openai/gpt-5.5 Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/openai~1gpt-5.5` | | | | | | | | 13 | [Claude Sonnet 5](https://benchize.com/models/claude-sonnet-5-max.md)Max | 89.56% | SE ±0.75 pp | $0.00957 | 9.31source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 89.555 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.009566 \| USD \| per_test \| Lower is better \| \| `latency` \| 9.309 \| 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: ±0.75 pp sourcePopulation 151 stderr 0.753 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-sonnet-5 Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/anthropic~1claude-sonnet-5` | | | | | | | | 14 | [DeepSeek V4 Pro 0813](https://benchize.com/models/deepseek-v4-pro-0813-max.md)Max | 89.55% | SE ±0.72 pp | $0.00357 | 17.96source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 89.548 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.003565 \| USD \| per_test \| Lower is better \| \| `latency` \| 17.96 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 384000 provider DeepSeek reasoning Not reported Reasoning effort max Temperature Not reported topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.72 pp sourcePopulation 151 stderr 0.715 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label deepseek/deepseek-v4-pro-0813 Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/deepseek~1deepseek-v4-pro-0813` | | | | | | | | 15 | [MiniMax M2.7](https://benchize.com/models/minimax-m2-7.md) | 89.34% | SE ±0.69 pp | $0.000458 | 8.91source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 89.339 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.000458 \| USD \| per_test \| Lower is better \| \| `latency` \| 8.912 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 196608 provider MiniMax reasoning Not reported Reasoning effort unknown Temperature 1 topP 0.95 verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.69 pp sourcePopulation 151 stderr 0.689 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label minimax/MiniMax-M2.7 Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/minimax~1MiniMax-M2.7` | | | | | | | | 16 | [MiniMax M3](https://benchize.com/models/minimax-m3.md) | 89.28% | SE ±0.74 pp | $0.00109 | 9.22source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 89.276 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.001085 \| USD \| per_test \| Lower is better \| \| `latency` \| 9.224 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 512000 provider MiniMax reasoning Not reported Reasoning effort unknown Temperature 1 topP 0.95 verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.74 pp sourcePopulation 151 stderr 0.742 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label minimax/MiniMax-M3 Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/minimax~1MiniMax-M3` | | | | | | | | 17 | [Gemini 3.1 Pro Preview](https://benchize.com/models/gemini-3-1-pro-preview-high.md)High | 89.26% | SE ±0.61 pp | $0.00593 | 10.23source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 89.257 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.005926 \| USD \| per_test \| Lower is better \| \| `latency` \| 10.23 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.61 pp sourcePopulation 151 stderr 0.611 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label google/gemini-3.1-pro-preview Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/google~1gemini-3.1-pro-preview` | | | | | | | | 18 | [DeepSeek V4 Pro](https://benchize.com/models/deepseek-v4-pro-max.md)Max | 89.12% | SE ±0.7 pp | $0.00431 | 115.98source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 89.117 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.004307 \| USD \| per_test \| Lower is better \| \| `latency` \| 115.976 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 256000 provider DeepSeek reasoning Not reported Reasoning effort max Temperature Not reported topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.7 pp sourcePopulation 151 stderr 0.704 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label deepseek/deepseek-v4-pro Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/deepseek~1deepseek-v4-pro` | | | | | | | | 19 | [DeepSeek V4.1 Flash](https://benchize.com/models/deepseek-v4-1-flash-high.md)High | 89.08% | SE ±0.78 pp | $0.00115 | 5.7source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 89.075 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.001151 \| USD \| per_test \| Lower is better \| \| `latency` \| 5.698 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 384000 provider DeepSeek reasoning Not reported Reasoning effort high Temperature Not reported topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.78 pp sourcePopulation 151 stderr 0.782 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label deepseek/deepseek-v4.1-flash Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/deepseek~1deepseek-v4.1-flash` | | | | | | | | 20 | [Claude Opus 4.8](https://benchize.com/models/claude-opus-4-8-max.md)Max | 88.94% | SE ±0.61 pp | $0.01 | 4.89source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 88.937 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.011109 \| USD \| per_test \| Lower is better \| \| `latency` \| 4.886 \| 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: ±0.61 pp sourcePopulation 151 stderr 0.61 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-opus-4-8 Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/anthropic~1claude-opus-4-8` | | | | | | | | 21 | [DeepSeek V4 Flash 0731](https://benchize.com/models/deepseek-v4-flash-0731-high.md)High | 88.89% | SE ±0.74 pp | $0.00036 | 3.2source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 88.892 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.00036 \| USD \| per_test \| Lower is better \| \| `latency` \| 3.197 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 256000 provider DeepSeek reasoning Not reported Reasoning effort high Temperature Not reported topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.74 pp sourcePopulation 151 stderr 0.742 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label deepseek/deepseek-v4-flash-0731 Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/deepseek~1deepseek-v4-flash-0731` | | | | | | | | 22 | [GPT-5.6 Luna](https://benchize.com/models/gpt-5-6-luna-max.md)Max | 88.86% | SE ±0.94 pp | $0.000524 | 3.78source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 88.864 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.000524 \| USD \| per_test \| Lower is better \| \| `latency` \| 3.783 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 128000 provider OpenAI reasoning Not reported Reasoning effort max Temperature Not reported topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.94 pp sourcePopulation 151 stderr 0.935 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label openai/gpt-5.6-luna Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/openai~1gpt-5.6-luna` | | | | | | | | 23 | [Mercury 2.5](https://benchize.com/models/mercury-2-5-high.md)High | 88.54% | SE ±0.84 pp | $0.000651 | 3.43source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 88.542 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.000651 \| USD \| per_test \| Lower is better \| \| `latency` \| 3.431 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Inception reasoning Not reported Reasoning effort high Temperature Not reported topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.84 pp sourcePopulation 151 stderr 0.844 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label inception/mercury-2.5 Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/inception~1mercury-2.5` | | | | | | | | 24 | [Gemini 3.1 Flash Lite Preview](https://benchize.com/models/gemini-3-1-flash-lite-preview-high.md)High | 88.47% | SE ±0.71 pp | $0.000289 | 2.9source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 88.472 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.000289 \| USD \| per_test \| Lower is better \| \| `latency` \| 2.904 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.71 pp sourcePopulation 151 stderr 0.714 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label google/gemini-3.1-flash-lite-preview Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/google~1gemini-3.1-flash-lite-preview` | | | | | | | | 25 | [Gemini 3.5 Flash-Lite](https://benchize.com/models/gemini-3-5-flash-lite-high.md)High | 88.4% | SE ±0.72 pp | $0.00131 | 1.78source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 88.404 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.001313 \| USD \| per_test \| Lower is better \| \| `latency` \| 1.784 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.72 pp sourcePopulation 151 stderr 0.721 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label google/gemini-3.5-flash-lite Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/google~1gemini-3.5-flash-lite` | | | | | | | | 26 | [MiMo-V2.5-Pro](https://benchize.com/models/mimo-v2-5-pro.md) | 88.03% | SE ±0.84 pp | $0.000694 | 21.46source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 88.031 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.000694 \| USD \| per_test \| Lower is better \| \| `latency` \| 21.462 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 128000 provider Xiaomi reasoning Not reported Reasoning effort unknown Temperature 1 topP 0.95 verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.84 pp sourcePopulation 151 stderr 0.842 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label xiaomi/mimo-v2.5-pro Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/xiaomi~1mimo-v2.5-pro` | | | | | | | | 27 | [Claude Sonnet 4.6](https://benchize.com/models/claude-sonnet-4-6-max.md)Max | 87.99% | SE ±1.02 pp | $0.00379 | 27.37source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 87.991 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.003786 \| USD \| per_test \| Lower is better \| \| `latency` \| 27.372 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning true Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.02 pp sourcePopulation 151 stderr 1.018 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-sonnet-4-6 Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/anthropic~1claude-sonnet-4-6` | | | | | | | | 28 | [Muse Spark 1.2](https://benchize.com/models/muse-spark-1-2-xhigh.md)Extra high | 87.49% | SE ±0.79 pp | $0.00731 | 36.04source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 87.485 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.00731 \| USD \| per_test \| Lower is better \| \| `latency` \| 36.037 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 131072 provider Meta reasoning Not reported Reasoning effort xhigh Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.79 pp sourcePopulation 151 stderr 0.794 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label meta/muse_spark_1_2 Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/meta~1muse_spark_1_2` | | | | | | | | 29 | [Nemotron 3 Ultra](https://benchize.com/models/nemotron-3-ultra-550b-a55b.md) | 87.47% | SE ±0.77 pp | — | 6.25source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 87.47 \| percent \| reported \| Higher is better \| \| `latency` \| 6.253 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 128000 provider Nvidia reasoning Not reported Reasoning effort unknown Temperature Not reported topP 0.95 verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.77 pp sourcePopulation 151 stderr 0.772 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label nvidia/nemotron-3-ultra-550b-a55b Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/nvidia~1nemotron-3-ultra-550b-a55b` | | | | | | | | 30 | [Muse Spark 1.1](https://benchize.com/models/muse-spark-1-1-xhigh.md)Extra high | 87.33% | SE ±0.74 pp | $0.00837 | 18.31source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 87.328 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.008371 \| USD \| per_test \| Lower is better \| \| `latency` \| 18.313 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 256000 provider Meta reasoning Not reported Reasoning effort xhigh Temperature Not reported topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.74 pp sourcePopulation 151 stderr 0.742 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label meta/muse_spark_1_1 Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/meta~1muse_spark_1_1` | | | | | | | | 31 | [GPT-4o (2024-11-20)](https://benchize.com/models/gpt-4o-2024-11-20-high.md)High | 87.21% | SE ±0.82 pp | — | 0.33source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 87.208 \| percent \| reported \| Higher is better \| \| `latency` \| 0.328 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 16384 provider OpenAI reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.82 pp sourcePopulation 151 stderr 0.818 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label openai/gpt-4o-2024-11-20 Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/openai~1gpt-4o-2024-11-20` | | | | | | | | 32 | [GPT-4o (2024-08-06)](https://benchize.com/models/gpt-4o-2024-08-06-high.md)High | 86.44% | SE ±0.77 pp | — | 0.34source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 86.442 \| percent \| reported \| Higher is better \| \| `latency` \| 0.343 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 16384 provider OpenAI reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.77 pp sourcePopulation 151 stderr 0.773 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label openai/gpt-4o-2024-08-06 Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/openai~1gpt-4o-2024-08-06` | | | | | | | | 33 | [MiMo-V2.5](https://benchize.com/models/mimo-v2-5.md) | 86.39% | SE ±0.84 pp | $0.000189 | 8.69source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 86.386 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.000189 \| USD \| per_test \| Lower is better \| \| `latency` \| 8.689 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 128000 provider Xiaomi reasoning Not reported Reasoning effort unknown Temperature 1 topP 0.95 verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.84 pp sourcePopulation 151 stderr 0.844 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label xiaomi/mimo-v2.5 Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/xiaomi~1mimo-v2.5` | | | | | | | | 34 | [MiniMax M2.5](https://benchize.com/models/minimax-m2-5.md) | 86.24% | SE ±0.9 pp | $0.00125 | 13.52source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 86.241 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.001249 \| USD \| per_test \| Lower is better \| \| `latency` \| 13.522 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 196608 provider MiniMax reasoning Not reported Reasoning effort unknown Temperature 1 topP 0.95 verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.9 pp sourcePopulation 151 stderr 0.903 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label minimax/MiniMax-M2.5 Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/minimax~1MiniMax-M2.5` | | | | | | | | 35 | [Laguna M 1](https://benchize.com/models/laguna-m-1.md) | 85.46% | SE ±0.94 pp | — | 45.3source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 85.459 \| percent \| reported \| Higher is better \| \| `latency` \| 45.298 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 32768 provider Poolside reasoning Not reported Reasoning effort unknown Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.94 pp sourcePopulation 151 stderr 0.941 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label poolside/laguna-m.1 Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/poolside~1laguna-m.1` | | | | | | | | 36 | [MiniMax M2.1](https://benchize.com/models/minimax-m2-1.md) | 85.32% | SE ±0.8 pp | $0.000695 | 9.64source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 85.318 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.000695 \| USD \| per_test \| Lower is better \| \| `latency` \| 9.637 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 131000 provider MiniMax reasoning Not reported Reasoning effort unknown Temperature 1 topP 0.95 verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.8 pp sourcePopulation 151 stderr 0.804 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label minimax/MiniMax-M2.1 Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/minimax~1MiniMax-M2.1` | | | | | | | | 37 | [Command A](https://benchize.com/models/command-a.md) | 83.08% | SE ±0.96 pp | — | 0.57source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 83.079 \| percent \| reported \| Higher is better \| \| `latency` \| 0.573 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 8000 provider Cohere reasoning Not reported Reasoning effort unknown Temperature 0.3 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.96 pp sourcePopulation 151 stderr 0.962 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label cohere/command-a-03-2025 Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/cohere~1command-a-03-2025` | | | | | | | | 38 | [GPT-4 Turbo](https://benchize.com/models/gpt-4-turbo-high.md)High | 82.55% | SE ±0.95 pp | — | 0.69source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 82.547 \| percent \| reported \| Higher is better \| \| `latency` \| 0.69 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 4096 provider OpenAI reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.95 pp sourcePopulation 151 stderr 0.954 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label openai/gpt-4-turbo Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/openai~1gpt-4-turbo` | | | | | | | | 39 | [Command A Plus](https://benchize.com/models/command-a-plus.md) | 82.12% | SE ±1.01 pp | $0.01 | 9.62source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 82.12 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.012736 \| USD \| per_test \| Lower is better \| \| `latency` \| 9.621 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 64000 provider Cohere reasoning Not reported Reasoning effort unknown Temperature 0.9 topP 0.95 verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.01 pp sourcePopulation 151 stderr 1.013 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label cohere/command-a-plus-05-2026 Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/cohere~1command-a-plus-05-2026` | | | | | | | | 40 | [Gemini 3.5 Flash](https://benchize.com/models/gemini-3-5-flash-high.md)High | 81.94% | SE ±1.06 pp | $0.00642 | 27.31source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 81.939 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.00642 \| USD \| per_test \| Lower is better \| \| `latency` \| 27.306 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65536 provider Google reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.06 pp sourcePopulation 151 stderr 1.055 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label google/gemini-3.5-flash Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/google~1gemini-3.5-flash` | | | | | | | | 41 | [Mistral Large 3 2512](https://benchize.com/models/mistral-large-2512.md) | 81.03% | SE ±1 pp | $0.000225 | 0.79source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 81.032 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.000225 \| USD \| per_test \| Lower is better \| \| `latency` \| 0.793 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 256000 provider Mistral AI reasoning Not reported Reasoning effort unknown Temperature 0.065 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1 pp sourcePopulation 151 stderr 1.001 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label mistralai/mistral-large-2512 Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/mistralai~1mistral-large-2512` | | | | | | | | 42 | [Gemini 2.5 Flash Lite Preview 09-2025](https://benchize.com/models/gemini-2-5-flash-lite-preview-09-2025.md) | 79.88% | SE ±1.05 pp | — | 0.36source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 79.878 \| percent \| reported \| Higher is better \| \| `latency` \| 0.359 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 65535 provider Google reasoning Not reported Reasoning effort unknown Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.05 pp sourcePopulation 151 stderr 1.051 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label google/gemini-2.5-flash-lite-preview-09-2025 Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/google~1gemini-2.5-flash-lite-preview-09-2025` | | | | | | | | 43 | [Laguna Xs 2](https://benchize.com/models/laguna-xs-2.md) | 77.1% | SE ±0.92 pp | — | 12.55source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 77.103 \| percent \| reported \| Higher is better \| \| `latency` \| 12.554 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 32768 provider Poolside reasoning Not reported Reasoning effort unknown Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.92 pp sourcePopulation 151 stderr 0.915 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label poolside/laguna-xs.2 Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/poolside~1laguna-xs.2` | | | | | | | | 44 | [GPT-3.5 Turbo](https://benchize.com/models/gpt-3-5-turbo-high.md)High | 67.4% | SE ±1.13 pp | — | 0.35source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 67.401 \| percent \| reported \| Higher is better \| \| `latency` \| 0.345 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 4096 provider OpenAI reasoning Not reported Reasoning effort high Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.13 pp sourcePopulation 151 stderr 1.13 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label openai/gpt-3.5-turbo Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/conclusion_tasks/openai~1gpt-3.5-turbo` | | | | | | | --- Cite as: LegalBench / conclusion tasks — Benchize, https://benchize.com/benchmarks/vals-ai%3Alegal_bench%3Aconclusion_tasks (retrieved 2026-10-21)