# Harvey's Legal Agent Benchmark v1 / criteria pass rate healthcare life sciences > Source: Benchize (https://benchize.com) — https://benchize.com/benchmarks/vals-ai%3Ahlab-v1%3Acriteria_pass_rate__healthcare_life_sciences. When citing or reusing this content, credit "Benchize" and link to https://benchize.com/benchmarks/vals-ai%3Ahlab-v1%3Acriteria_pass_rate__healthcare_life_sciences. Tests an agent's ability to complete legal work using documents, spreadsheets, presentations, and file-system tools Canonical URL: https://benchize.com/benchmarks/vals-ai%3Ahlab-v1%3Acriteria_pass_rate__healthcare_life_sciences 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. Tests an agent's ability to complete legal work using documents, spreadsheets, presentations, and file-system tools Professional work · Models tested: 49 · Score publishers: Vals AI ## Model results | # | Model | Score | ± | Cost | Latency | | | --- | --- | --- | --- | --- | --- | --- | | 1 | [Muse Spark 1.2](https://benchize.com/models/muse-spark-1-2-xhigh.md)Extra high | 95.38% | SE ±1.21 pp | $2.6 | 1,815.46source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 95.38 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 2.602242 \| USD \| per_test \| Lower is better \| \| `latency` \| 1815.46 \| 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: ±1.21 pp sourcePopulation 81 stderr 1.208 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 meta/muse_spark_1_2 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/meta~1muse_spark_1_2` | | | | | | | | 2 | [Muse Wonder](https://benchize.com/models/muse-wonder-max.md)Max | 94.06% | SE ±1.36 pp | $5.63 | 1,364.17source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 94.059 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 5.625299 \| USD \| per_test \| Lower is better \| \| `latency` \| 1364.174 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 131072 provider Meta reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.36 pp sourcePopulation 81 stderr 1.36 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 meta/muse-wonder Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/meta~1muse-wonder` | | | | | | | | 3 | [Muse Spark 1.3](https://benchize.com/models/muse-spark-1-3-xhigh.md)Extra high | 93.73% | SE ±1.4 pp | $1.87 | 909.88source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 93.729 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 1.865298 \| USD \| per_test \| Lower is better \| \| `latency` \| 909.876 \| 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: ±1.4 pp sourcePopulation 81 stderr 1.395 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 meta/muse_spark_1_3 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/meta~1muse_spark_1_3` | | | | | | | | 3 | [Gemini 3.8 Flash](https://benchize.com/models/gemini-3-8-flash-high.md)High | 93.73% | SE ±1.4 pp | $3.47 | 833.11source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 93.729 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 3.474289 \| USD \| per_test \| Lower is better \| \| `latency` \| 833.11 \| 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: ±1.4 pp sourcePopulation 81 stderr 1.395 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 google/gemini-3.8-flash Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/google~1gemini-3.8-flash` | | | | | | | | 5 | [Muse Spark 1.1](https://benchize.com/models/muse-spark-1-1-xhigh.md)Extra high | 92.74% | SE ±1.49 pp | $0.99 | 1,001.37source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 92.739 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.991405 \| USD \| per_test \| Lower is better \| \| `latency` \| 1001.374 \| 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: ±1.49 pp sourcePopulation 81 stderr 1.493 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 meta/muse_spark_1_1 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/meta~1muse_spark_1_1` | | | | | | | | 5 | [Gemini 4 Argon](https://benchize.com/models/gemini-4-argon-high.md)High | 92.74% | SE ±1.49 pp | $8.84 | 2,917.03source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 92.739 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 8.843854 \| USD \| per_test \| Lower is better \| \| `latency` \| 2917.028 \| 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: ±1.49 pp sourcePopulation 81 stderr 1.493 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 google/gemini-4-argon Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/google~1gemini-4-argon` | | | | | | | | 7 | [MiMo-V2.6-Flash](https://benchize.com/models/mimo-v2-6-flash.md) | 92.41% | SE ±1.52 pp | $0.09 | 1,607.47source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 92.409 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.09248 \| USD \| per_test \| Lower is better \| \| `latency` \| 1607.468 \| 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: ±1.52 pp sourcePopulation 81 stderr 1.524 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 xiaomi/mimo-v2.6-flash Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/xiaomi~1mimo-v2.6-flash` | | | | | | | | 8 | [MiMo-V2.6-Pro](https://benchize.com/models/mimo-v2-6-pro.md) | 92.08% | SE ±1.55 pp | $0.21 | 2,184.58source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 92.079 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.21495 \| USD \| per_test \| Lower is better \| \| `latency` \| 2184.576 \| 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: ±1.55 pp sourcePopulation 81 stderr 1.554 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 xiaomi/mimo-v2.6-pro Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/xiaomi~1mimo-v2.6-pro` | | | | | | | | 9 | [Gemini 3.7 Flash](https://benchize.com/models/gemini-3-7-flash-high.md)High | 91.75% | SE ±1.58 pp | $3 | 511.32source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 91.749 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 3.002272 \| USD \| per_test \| Lower is better \| \| `latency` \| 511.32 \| 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: ±1.58 pp sourcePopulation 81 stderr 1.583 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 google/gemini-3.7-flash Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/google~1gemini-3.7-flash` | | | | | | | | 10 | [Claude Opus 4.8](https://benchize.com/models/claude-opus-4-8-max.md)Max | 91.42% | SE ±1.61 pp | $10.48 | 1,343.57source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 91.419 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 10.479166 \| USD \| per_test \| Lower is better \| \| `latency` \| 1343.57 \| 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: ±1.61 pp sourcePopulation 81 stderr 1.612 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-opus-4-8 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/anthropic~1claude-opus-4-8` | | | | | | | | 10 | [Claude Fable 5](https://benchize.com/models/claude-fable-5-max.md)Max | 91.42% | SE ±1.61 pp | $23.08 | 1,645.03source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 91.419 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 23.076834 \| USD \| per_test \| Lower is better \| \| `latency` \| 1645.034 \| 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: ±1.61 pp sourcePopulation 81 stderr 1.612 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__healthcare_life_sciences/anthropic~1claude-fable-5` | | | | | | | | 12 | [DeepSeek V4 Flash 0731](https://benchize.com/models/deepseek-v4-flash-0731-high.md)High | 91.09% | SE ±1.64 pp | $0.06 | 1,043.86source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 91.089 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.056365 \| USD \| per_test \| Lower is better \| \| `latency` \| 1043.858 \| 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: ±1.64 pp sourcePopulation 81 stderr 1.639 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 deepseek/deepseek-v4-flash-0731 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/deepseek~1deepseek-v4-flash-0731` | | | | | | | | 13 | [DeepSeek V4 Pro 0813](https://benchize.com/models/deepseek-v4-pro-0813-max.md)Max | 90.76% | SE ±1.67 pp | $0.16 | 1,207.57source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 90.759 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.164353 \| USD \| per_test \| Lower is better \| \| `latency` \| 1207.572 \| 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: ±1.67 pp sourcePopulation 81 stderr 1.666 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 deepseek/deepseek-v4-pro-0813 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/deepseek~1deepseek-v4-pro-0813` | | | | | | | | 13 | [Claude Opus 5](https://benchize.com/models/claude-opus-5-max.md)Max | 90.76% | SE ±1.67 pp | $24.92 | 3,583.41source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 90.759 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 24.921853 \| USD \| per_test \| Lower is better \| \| `latency` \| 3583.414 \| 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: ±1.67 pp sourcePopulation 81 stderr 1.666 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-opus-5 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/anthropic~1claude-opus-5` | | | | | | | | 15 | [DeepSeek V4.1 Flash](https://benchize.com/models/deepseek-v4-1-flash-max.md)Max | 90.43% | SE ±1.69 pp | $0.17 | 622.19source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 90.429 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.166133 \| USD \| per_test \| Lower is better \| \| `latency` \| 622.19 \| 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 1 topP 1 verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.69 pp sourcePopulation 81 stderr 1.693 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 deepseek/deepseek-v4.1-flash Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/deepseek~1deepseek-v4.1-flash` | | | | | | | | 16 | [Step 5 Preview](https://benchize.com/models/step-5-preview.md) | 90.1% | SE ±1.72 pp | $1.29 | 2,200.13source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 90.099 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 1.289815 \| USD \| per_test \| Lower is better \| \| `latency` \| 2200.132 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 1024000 provider Stepfun reasoning Not reported Reasoning effort unknown Temperature 1 topP 0.95 verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.72 pp sourcePopulation 81 stderr 1.719 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 stepfun/step-5-preview Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/stepfun~1step-5-preview` | | | | | | | | 17 | [MiniMax M3](https://benchize.com/models/minimax-m3.md) | 89.77% | SE ±1.74 pp | $1.46 | 1,436.11source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 89.769 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 1.4604 \| USD \| per_test \| Lower is better \| \| `latency` \| 1436.112 \| 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: ±1.74 pp sourcePopulation 81 stderr 1.744 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 minimax/MiniMax-M3 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/minimax~1MiniMax-M3` | | | | | | | | 18 | [DeepSeek V4.1 Flash 1017](https://benchize.com/models/deepseek-v4-1-flash-1017-max.md)Max | 89.44% | SE ±1.77 pp | $0.51 | 1,310.75source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 89.439 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.506684 \| USD \| per_test \| Lower is better \| \| `latency` \| 1310.752 \| 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 1 topP 1 verbosity Not reported ### Reported statistics Uncertainty Standard error: ±1.77 pp sourcePopulation 81 stderr 1.769 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 deepseek/deepseek-v4.1-flash-1017 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/deepseek~1deepseek-v4.1-flash-1017` | | | | | | | | 19 | [Claude Opus 4.7](https://benchize.com/models/claude-opus-4-7-max.md)Max | 89.11% | SE ±1.79 pp | $14.12 | 1,954.01source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 89.109 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 14.122835 \| USD \| per_test \| Lower is better \| \| `latency` \| 1954.012 \| 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: ±1.79 pp sourcePopulation 81 stderr 1.793 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-opus-4-7 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/anthropic~1claude-opus-4-7` | | | | | | | | 20 | [Claude Sonnet 4.6](https://benchize.com/models/claude-sonnet-4-6-max.md)Max | 88.78% | SE ±1.82 pp | $3.82 | 1,277.57source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 88.779 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 3.819729 \| USD \| per_test \| Lower is better \| \| `latency` \| 1277.572 \| 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: ±1.82 pp sourcePopulation 81 stderr 1.816 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-sonnet-4-6 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/anthropic~1claude-sonnet-4-6` | | | | | | | | 21 | [Claude Fable 5.1](https://benchize.com/models/claude-fable-5-1-max.md)Max | 88.45% | SE ±1.84 pp | $41.98 | 4,427.73source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 88.449 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 41.981832 \| USD \| per_test \| Lower is better \| \| `latency` \| 4427.734 \| 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: ±1.84 pp sourcePopulation 81 stderr 1.839 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-5-1 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/anthropic~1claude-fable-5-1` | | | | | | | | 22 | [MiMo-V2.5-Pro](https://benchize.com/models/mimo-v2-5-pro.md) | 87.79% | SE ±1.88 pp | $0.12 | 541.79source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 87.789 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.121036 \| USD \| per_test \| Lower is better \| \| `latency` \| 541.786 \| 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: ±1.88 pp sourcePopulation 81 stderr 1.884 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 xiaomi/mimo-v2.5-pro Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/xiaomi~1mimo-v2.5-pro` | | | | | | | | 23 | [Gemini 3.6 Flash](https://benchize.com/models/gemini-3-6-flash-high.md)High | 87.13% | SE ±1.93 pp | $1.9 | 1,131.51source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 87.129 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 1.902096 \| USD \| per_test \| Lower is better \| \| `latency` \| 1131.514 \| 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: ±1.93 pp sourcePopulation 81 stderr 1.927 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 google/gemini-3.6-flash Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/google~1gemini-3.6-flash` | | | | | | | | 23 | [GPT-6.1 Sol](https://benchize.com/models/gpt-6-1-sol-max.md)Max | 87.13% | SE ±1.93 pp | $4.34 | 3,718.03source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 87.129 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 4.342941 \| USD \| per_test \| Lower is better \| \| `latency` \| 3718.032 \| 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: ±1.93 pp sourcePopulation 81 stderr 1.927 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 openai/gpt-6.1-sol Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/openai~1gpt-6.1-sol` | | | | | | | | 23 | [Claude Sonnet 5.5](https://benchize.com/models/claude-sonnet-5-5-max.md)Max | 87.13% | SE ±1.93 pp | $19.89 | 4,075.73source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 87.129 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 19.894952 \| USD \| per_test \| Lower is better \| \| `latency` \| 4075.734 \| 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: ±1.93 pp sourcePopulation 81 stderr 1.927 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-sonnet-5-5 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/anthropic~1claude-sonnet-5-5` | | | | | | | | 26 | [GPT-6 Astra](https://benchize.com/models/gpt-6-astra-max.md)Max | 86.47% | SE ±1.97 pp | $20.84 | 1,509.92source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 86.469 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 20.8409 \| USD \| per_test \| Lower is better \| \| `latency` \| 1509.924 \| 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: ±1.97 pp sourcePopulation 81 stderr 1.968 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 openai/gpt-6-astra Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/openai~1gpt-6-astra` | | | | | | | | 27 | [Claude Opus 5.5](https://benchize.com/models/claude-opus-5-5-max.md)Max | 86.14% | SE ±1.99 pp | $28.86 | 4,054.91source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 86.139 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 28.864668 \| USD \| per_test \| Lower is better \| \| `latency` \| 4054.91 \| 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: ±1.99 pp sourcePopulation 81 stderr 1.988 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-opus-5-5 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/anthropic~1claude-opus-5-5` | | | | | | | | 27 | [DeepSeek V4 Pro](https://benchize.com/models/deepseek-v4-pro-max.md)Max | 86.14% | SE ±1.99 pp | $0.85 | 435.38source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 86.139 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.848233 \| USD \| per_test \| Lower is better \| \| `latency` \| 435.376 \| 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: ±1.99 pp sourcePopulation 81 stderr 1.988 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 deepseek/deepseek-v4-pro Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/deepseek~1deepseek-v4-pro` | | | | | | | | 29 | [Claude Sonnet 5](https://benchize.com/models/claude-sonnet-5-max.md)Max | 85.48% | SE ±2.03 pp | $8.62 | 2,340.06source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 85.479 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 8.616817 \| USD \| per_test \| Lower is better \| \| `latency` \| 2340.06 \| 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: ±2.03 pp sourcePopulation 81 stderr 2.027 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-sonnet-5 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/anthropic~1claude-sonnet-5` | | | | | | | | 30 | [GPT-5.6 Sol](https://benchize.com/models/gpt-5-6-sol-max.md)Max | 82.51% | SE ±2.19 pp | $11.6 | 1,594.3source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 82.508 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 11.601398 \| USD \| per_test \| Lower is better \| \| `latency` \| 1594.302 \| 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: ±2.19 pp sourcePopulation 81 stderr 2.186 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 openai/gpt-5.6-sol Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/openai~1gpt-5.6-sol` | | | | | | | | 31 | [GPT-5.5](https://benchize.com/models/gpt-5-5-xhigh.md)Extra high | 81.85% | SE ±2.22 pp | $4.01 | 812.9source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 81.848 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 4.010513 \| USD \| per_test \| Lower is better \| \| `latency` \| 812.896 \| 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: ±2.22 pp sourcePopulation 81 stderr 2.218 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 openai/gpt-5.5 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/openai~1gpt-5.5` | | | | | | | | 32 | [GPT-6 Luna](https://benchize.com/models/gpt-6-luna-max.md)Max | 80.86% | SE ±2.26 pp | $0.24 | 997.66source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 80.858 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.239122 \| USD \| per_test \| Lower is better \| \| `latency` \| 997.662 \| 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: ±2.26 pp sourcePopulation 81 stderr 2.264 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 openai/gpt-6-luna Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/openai~1gpt-6-luna` | | | | | | | | 33 | [Gemini 3.5 Flash](https://benchize.com/models/gemini-3-5-flash-high.md)High | 80.53% | SE ±2.28 pp | $2.41 | 575.66source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 80.528 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 2.41183 \| USD \| per_test \| Lower is better \| \| `latency` \| 575.66 \| 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: ±2.28 pp sourcePopulation 81 stderr 2.279 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 google/gemini-3.5-flash Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/google~1gemini-3.5-flash` | | | | | | | | 33 | [MiMo-V2.5](https://benchize.com/models/mimo-v2-5.md) | 80.53% | SE ±2.28 pp | $0.04 | 444.82source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 80.528 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.043026 \| USD \| per_test \| Lower is better \| \| `latency` \| 444.818 \| 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: ±2.28 pp sourcePopulation 81 stderr 2.279 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 xiaomi/mimo-v2.5 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/xiaomi~1mimo-v2.5` | | | | | | | | 35 | [GPT-6 Sol](https://benchize.com/models/gpt-6-sol-max.md)Max | 79.87% | SE ±2.31 pp | $2.39 | 802.49source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 79.868 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 2.391775 \| USD \| per_test \| Lower is better \| \| `latency` \| 802.494 \| 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: ±2.31 pp sourcePopulation 81 stderr 2.307 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 openai/gpt-6-sol Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/openai~1gpt-6-sol` | | | | | | | | 36 | [GPT-5.6 Terra](https://benchize.com/models/gpt-5-6-terra-max.md)Max | 78.88% | SE ±2.35 pp | $3.2 | 965.99source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 78.878 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 3.200117 \| USD \| per_test \| Lower is better \| \| `latency` \| 965.994 \| 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: ±2.35 pp sourcePopulation 81 stderr 2.349 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 openai/gpt-5.6-terra Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/openai~1gpt-5.6-terra` | | | | | | | | 37 | [MiniMax M2.7](https://benchize.com/models/minimax-m2-7.md) | 78.55% | SE ±2.36 pp | $0.86 | 485.53source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 78.548 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.860793 \| USD \| per_test \| Lower is better \| \| `latency` \| 485.53 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 80000 provider MiniMax reasoning Not reported Reasoning effort unknown Temperature 1 topP 0.95 verbosity Not reported ### Reported statistics Uncertainty Standard error: ±2.36 pp sourcePopulation 81 stderr 2.362 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 minimax/MiniMax-M2.7 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/minimax~1MiniMax-M2.7` | | | | | | | | 38 | [GPT-5.6 Luna](https://benchize.com/models/gpt-5-6-luna-max.md)Max | 78.22% | SE ±2.38 pp | $0.29 | 702.1source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 78.218 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.289523 \| USD \| per_test \| Lower is better \| \| `latency` \| 702.102 \| 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: ±2.38 pp sourcePopulation 81 stderr 2.375 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 openai/gpt-5.6-luna Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/openai~1gpt-5.6-luna` | | | | | | | | 39 | [Nemotron 3 Ultra](https://benchize.com/models/nemotron-3-ultra-550b-a55b.md) | 69.64% | SE ±2.65 pp | — | 1,443.01source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 69.637 \| percent \| reported \| Higher is better \| \| `latency` \| 1443.008 \| 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 1 topP 0.95 verbosity Not reported ### Reported statistics Uncertainty Standard error: ±2.65 pp sourcePopulation 81 stderr 2.646 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 nvidia/nemotron-3-ultra-550b-a55b Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/nvidia~1nemotron-3-ultra-550b-a55b` | | | | | | | | 40 | [Claude Haiku 5.5](https://benchize.com/models/claude-haiku-5-5-max.md)Max | 66.53% | SE ±2.98 pp | $3.4 | 2,395.06source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 66.534 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 3.397341 \| USD \| per_test \| Lower is better \| \| `latency` \| 2395.058 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature Not reported topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±2.98 pp sourcePopulation 81 stderr 2.984 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-haiku-5-5 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/anthropic~1claude-haiku-5-5` | | | | | | | | 41 | [Gemini 3.1 Pro Preview](https://benchize.com/models/gemini-3-1-pro-preview-high.md)High | 60.4% | SE ±2.81 pp | $0.74 | 597.64source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 60.396 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.741325 \| USD \| per_test \| Lower is better \| \| `latency` \| 597.642 \| 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: ±2.81 pp sourcePopulation 81 stderr 2.814 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 google/gemini-3.1-pro-preview Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/google~1gemini-3.1-pro-preview` | | | | | | | | 41 | [Gemini 3.5 Flash-Lite](https://benchize.com/models/gemini-3-5-flash-lite-high.md)High | 60.4% | SE ±2.81 pp | $0.28 | 141.4source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 60.396 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.284177 \| USD \| per_test \| Lower is better \| \| `latency` \| 141.404 \| 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: ±2.81 pp sourcePopulation 81 stderr 2.814 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 google/gemini-3.5-flash-lite Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/google~1gemini-3.5-flash-lite` | | | | | | | | 43 | [Laguna M 1](https://benchize.com/models/laguna-m-1.md) | 60.07% | SE ±2.82 pp | $0.57 | 511.99source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 60.066 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.572801 \| USD \| per_test \| Lower is better \| \| `latency` \| 511.986 \| 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: ±2.82 pp sourcePopulation 81 stderr 2.818 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 poolside/laguna-m.1 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/poolside~1laguna-m.1` | | | | | | | | 44 | [Gemini 3 Flash Preview](https://benchize.com/models/gemini-3-flash-preview-high.md)High | 59.08% | SE ±2.83 pp | $0.29 | 215.34source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 59.076 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.288454 \| USD \| per_test \| Lower is better \| \| `latency` \| 215.342 \| 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: ±2.83 pp sourcePopulation 81 stderr 2.829 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 google/gemini-3-flash-preview Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/google~1gemini-3-flash-preview` | | | | | | | | 45 | [Mistral Medium 3.5](https://benchize.com/models/mistral-medium-3-5-high.md)High | 55.45% | SE ±2.86 pp | $6.26 | 4,695.76source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 55.446 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 6.264992 \| USD \| per_test \| Lower is better \| \| `latency` \| 4695.758 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 80000 provider Mistral AI reasoning Not reported Reasoning effort high Temperature 0.7 topP 0.95 verbosity Not reported ### Reported statistics Uncertainty Standard error: ±2.86 pp sourcePopulation 81 stderr 2.86 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 mistralai/mistral-medium-3.5 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/mistralai~1mistral-medium-3.5` | | | | | | | | 46 | [Laguna Xs 2](https://benchize.com/models/laguna-xs-2.md) | 42.9% | SE ±2.85 pp | $0.34 | 1,426.68source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 42.904 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.336544 \| USD \| per_test \| Lower is better \| \| `latency` \| 1426.678 \| 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: ±2.85 pp sourcePopulation 81 stderr 2.848 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 poolside/laguna-xs.2 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/poolside~1laguna-xs.2` | | | | | | | | 47 | [Mercury 2.5](https://benchize.com/models/mercury-2-5-high.md)High | 42.57% | SE ±2.85 pp | $0.21 | 112.73source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 42.574 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.210725 \| USD \| per_test \| Lower is better \| \| `latency` \| 112.726 \| 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: ±2.85 pp sourcePopulation 81 stderr 2.845 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 inception/mercury-2.5 Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/inception~1mercury-2.5` | | | | | | | | 48 | [Muse Spark Open](https://benchize.com/models/muse-spark-open-xhigh.md)Extra high | 34.98% | SE ±2.74 pp | $5.03 | 7,857.63source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 34.983 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 5.029791 \| USD \| per_test \| Lower is better \| \| `latency` \| 7857.632 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort Not reported harness Not reported maxOutputTokens 131072 provider Together AI reasoning Not reported Reasoning effort xhigh Temperature 1 topP 1 verbosity Not reported ### Reported statistics Uncertainty Standard error: ±2.74 pp sourcePopulation 81 stderr 2.744 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 meta/muse-spark-open Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/meta~1muse-spark-open` | | | | | | | | 49 | [Gemini 3.1 Flash Lite Preview](https://benchize.com/models/gemini-3-1-flash-lite-preview-high.md)High | 30.03% | SE ±2.64 pp | $0.05 | 46.9source duration | | | ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 30.033 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.049241 \| USD \| per_test \| Lower is better \| \| `latency` \| 46.904 \| 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: ±2.64 pp sourcePopulation 81 stderr 2.638 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 google/gemini-3.1-flash-lite-preview Source version 1 Observed 2026-10-21T16:45:53.909Z Evidence SHA-256 `65e7cd51e83066a42c9d782635dba5f2e9e1b189bd8118d3510cd17d2ddcba3f` Source locator `/tasks/criteria_pass_rate__healthcare_life_sciences/google~1gemini-3.1-flash-lite-preview` | | | | | | | --- Cite as: Harvey's Legal Agent Benchmark v1 / criteria pass rate healthcare life sciences — Benchize, https://benchize.com/benchmarks/vals-ai%3Ahlab-v1%3Acriteria_pass_rate__healthcare_life_sciences (retrieved 2026-10-21)