# LegalBench / interpretation tasks

> Source: Benchize (https://benchize.com) — https://benchize.com/benchmarks/vals-ai%3Alegal_bench%3Ainterpretation_tasks. When citing or reusing this content, credit "Benchize" and link to https://benchize.com/benchmarks/vals-ai%3Alegal_bench%3Ainterpretation_tasks.

Evaluating language models on a wide range of open source legal reasoning tasks.

Canonical URL: https://benchize.com/benchmarks/vals-ai%3Alegal_bench%3Ainterpretation_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 | [Gemini 3.5 Flash](https://benchize.com/models/gemini-3-5-flash-high.md)High | 88.68% | SE ±0.29 pp | $0.0062 | 30.37source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 88.675 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.006199 \| USD \| per_test \| Lower is better \| \| `latency` \| 30.374 \| 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: ±0.29 pp sourcePopulation 151 stderr 0.289 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/interpretation_tasks/google~1gemini-3.5-flash` |  |  |  |  |  |  |
| 2 | [Claude Fable 5.1](https://benchize.com/models/claude-fable-5-1-max.md)Max | 86.9% | SE ±0.19 pp | $0.04 | 10.01source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 86.902 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.043747 \| USD \| per_test \| Lower is better \| \| `latency` \| 10.008 \| 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: ±0.19 pp sourcePopulation 151 stderr 0.19 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/interpretation_tasks/anthropic~1claude-fable-5-1` |  |  |  |  |  |  |
| 3 | [Claude Fable 5](https://benchize.com/models/claude-fable-5-max.md)Max | 86.86% | SE ±0.16 pp | $0.03 | 8.12source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 86.863 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.028526 \| USD \| per_test \| Lower is better \| \| `latency` \| 8.119 \| source_duration \| reported_duration \| Lower is better \| ### Configuration computeEffort max harness Not reported maxOutputTokens 128000 provider Anthropic reasoning Not reported Reasoning effort max Temperature 1 topP Not reported verbosity Not reported ### Reported statistics Uncertainty Standard error: ±0.16 pp sourcePopulation 151 stderr 0.163 tokenTotals Not reported ### Source [Original source](https://www.vals.ai/_astro/benchmark_view_legal_bench.D6uTgA3b.json) Source Vals AI Original model label anthropic/claude-fable-5 Source version 1 Observed 2026-10-21T16:45:53.390Z Evidence SHA-256 `c8486ba756baa5cd3955d035edff67440e8db23e69d62ad1a9e18d15b8c4bc52` Source locator `/tasks/interpretation_tasks/anthropic~1claude-fable-5` |  |  |  |  |  |  |
| 4 | [Claude Opus 5](https://benchize.com/models/claude-opus-5-max.md)Max | 85.84% | SE ±0.19 pp | $0.01 | 3.08source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 85.84 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.010305 \| USD \| per_test \| Lower is better \| \| `latency` \| 3.075 \| 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: ±0.19 pp sourcePopulation 151 stderr 0.191 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/interpretation_tasks/anthropic~1claude-opus-5` |  |  |  |  |  |  |
| 5 | [Gemini 3.7 Flash](https://benchize.com/models/gemini-3-7-flash-high.md)High | 84.61% | SE ±0.35 pp | $0.0043 | 2.15source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 84.608 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.004299 \| USD \| per_test \| Lower is better \| \| `latency` \| 2.145 \| 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: ±0.35 pp sourcePopulation 151 stderr 0.348 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/interpretation_tasks/google~1gemini-3.7-flash` |  |  |  |  |  |  |
| 6 | [Gemini 3.8 Flash](https://benchize.com/models/gemini-3-8-flash-high.md)High | 84.59% | SE ±0.33 pp | $0.01 | 4.63source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 84.589 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.012231 \| USD \| per_test \| Lower is better \| \| `latency` \| 4.633 \| 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: ±0.33 pp sourcePopulation 151 stderr 0.325 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/interpretation_tasks/google~1gemini-3.8-flash` |  |  |  |  |  |  |
| 7 | [GPT-5.6 Sol](https://benchize.com/models/gpt-5-6-sol-max.md)Max | 84.4% | SE ±0.2 pp | $0.02 | 7.67source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 84.398 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.01698 \| USD \| per_test \| Lower is better \| \| `latency` \| 7.666 \| 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: ±0.2 pp sourcePopulation 151 stderr 0.203 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/interpretation_tasks/openai~1gpt-5.6-sol` |  |  |  |  |  |  |
| 8 | [Gemini 4 Argon](https://benchize.com/models/gemini-4-argon-high.md)High | 84.1% | SE ±0.21 pp | $0.02 | 12.5source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 84.097 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.016184 \| USD \| per_test \| Lower is better \| \| `latency` \| 12.499 \| 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: ±0.21 pp sourcePopulation 151 stderr 0.213 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/interpretation_tasks/google~1gemini-4-argon` |  |  |  |  |  |  |
| 9 | [Gemini 3 Flash Preview](https://benchize.com/models/gemini-3-flash-preview-high.md)High | 83.91% | SE ±0.3 pp | $0.00258 | 5.67source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 83.908 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.002582 \| USD \| per_test \| Lower is better \| \| `latency` \| 5.669 \| 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: ±0.3 pp sourcePopulation 151 stderr 0.302 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/interpretation_tasks/google~1gemini-3-flash-preview` |  |  |  |  |  |  |
| 10 | [Gemini 3.6 Flash](https://benchize.com/models/gemini-3-6-flash-high.md)High | 83.89% | SE ±0.33 pp | $0.00492 | 2.98source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 83.894 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.004921 \| USD \| per_test \| Lower is better \| \| `latency` \| 2.98 \| 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: ±0.33 pp sourcePopulation 151 stderr 0.329 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/interpretation_tasks/google~1gemini-3.6-flash` |  |  |  |  |  |  |
| 11 | [GPT-5.5](https://benchize.com/models/gpt-5-5-xhigh.md)Extra high | 83.84% | SE ±0.19 pp | $0.01 | 15.56source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 83.836 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.014325 \| USD \| per_test \| Lower is better \| \| `latency` \| 15.562 \| 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: ±0.19 pp sourcePopulation 151 stderr 0.194 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/interpretation_tasks/openai~1gpt-5.5` |  |  |  |  |  |  |
| 12 | [Gemini 3.1 Pro Preview](https://benchize.com/models/gemini-3-1-pro-preview-high.md)High | 83.82% | SE ±0.37 pp | $0.00609 | 9.14source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 83.823 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.006094 \| USD \| per_test \| Lower is better \| \| `latency` \| 9.143 \| 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: ±0.37 pp sourcePopulation 151 stderr 0.367 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/interpretation_tasks/google~1gemini-3.1-pro-preview` |  |  |  |  |  |  |
| 13 | [Claude Opus 4.7](https://benchize.com/models/claude-opus-4-7-max.md)Max | 83.61% | SE ±0.2 pp | $0.0077 | 24.39source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 83.609 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.007695 \| USD \| per_test \| Lower is better \| \| `latency` \| 24.393 \| 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: ±0.2 pp sourcePopulation 151 stderr 0.204 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/interpretation_tasks/anthropic~1claude-opus-4-7` |  |  |  |  |  |  |
| 14 | [MiniMax M3](https://benchize.com/models/minimax-m3.md) | 83.51% | SE ±0.2 pp | $0.00083 | 6.58source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 83.511 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.00083 \| USD \| per_test \| Lower is better \| \| `latency` \| 6.584 \| 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: ±0.2 pp sourcePopulation 151 stderr 0.199 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/interpretation_tasks/minimax~1MiniMax-M3` |  |  |  |  |  |  |
| 15 | [GPT-5.6 Terra](https://benchize.com/models/gpt-5-6-terra-max.md)Max | 83.18% | SE ±0.19 pp | $0.00242 | 2.35source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 83.178 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.002419 \| USD \| per_test \| Lower is better \| \| `latency` \| 2.352 \| 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: ±0.19 pp sourcePopulation 151 stderr 0.186 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/interpretation_tasks/openai~1gpt-5.6-terra` |  |  |  |  |  |  |
| 16 | [Muse Spark 1.1](https://benchize.com/models/muse-spark-1-1-xhigh.md)Extra high | 82.88% | SE ±0.2 pp | $0.00777 | 14.79source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 82.876 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.00777 \| USD \| per_test \| Lower is better \| \| `latency` \| 14.792 \| 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: ±0.2 pp sourcePopulation 151 stderr 0.199 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/interpretation_tasks/meta~1muse_spark_1_1` |  |  |  |  |  |  |
| 17 | [Claude Opus 4.8](https://benchize.com/models/claude-opus-4-8-max.md)Max | 82.7% | SE ±0.2 pp | $0.00753 | 2.25source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 82.703 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.007534 \| USD \| per_test \| Lower is better \| \| `latency` \| 2.245 \| 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: ±0.2 pp sourcePopulation 151 stderr 0.202 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/interpretation_tasks/anthropic~1claude-opus-4-8` |  |  |  |  |  |  |
| 18 | [Muse Spark 1.2](https://benchize.com/models/muse-spark-1-2-xhigh.md)Extra high | 82.54% | SE ±0.2 pp | $0.00536 | 27.91source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 82.542 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.005359 \| USD \| per_test \| Lower is better \| \| `latency` \| 27.908 \| 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: ±0.2 pp sourcePopulation 151 stderr 0.197 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/interpretation_tasks/meta~1muse_spark_1_2` |  |  |  |  |  |  |
| 19 | [Claude Sonnet 4.6](https://benchize.com/models/claude-sonnet-4-6-max.md)Max | 82.47% | SE ±0.19 pp | $0.00289 | 19.33source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 82.469 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.002893 \| USD \| per_test \| Lower is better \| \| `latency` \| 19.328 \| 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: ±0.19 pp sourcePopulation 151 stderr 0.191 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/interpretation_tasks/anthropic~1claude-sonnet-4-6` |  |  |  |  |  |  |
| 20 | [GPT-5.6 Luna](https://benchize.com/models/gpt-5-6-luna-max.md)Max | 82.14% | SE ±0.2 pp | $0.000696 | 4.46source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 82.142 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.000696 \| USD \| per_test \| Lower is better \| \| `latency` \| 4.461 \| 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: ±0.2 pp sourcePopulation 151 stderr 0.199 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/interpretation_tasks/openai~1gpt-5.6-luna` |  |  |  |  |  |  |
| 21 | [Mercury 2.5](https://benchize.com/models/mercury-2-5-high.md)High | 81.91% | SE ±0.2 pp | $0.000802 | 3.74source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 81.911 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.000802 \| USD \| per_test \| Lower is better \| \| `latency` \| 3.739 \| 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: ±0.2 pp sourcePopulation 151 stderr 0.201 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/interpretation_tasks/inception~1mercury-2.5` |  |  |  |  |  |  |
| 22 | [MiniMax M2.7](https://benchize.com/models/minimax-m2-7.md) | 81.65% | SE ±0.2 pp | $0.000625 | 8.58source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 81.653 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.000625 \| USD \| per_test \| Lower is better \| \| `latency` \| 8.579 \| 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: ±0.2 pp sourcePopulation 151 stderr 0.203 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/interpretation_tasks/minimax~1MiniMax-M2.7` |  |  |  |  |  |  |
| 23 | [Claude Sonnet 5](https://benchize.com/models/claude-sonnet-5-max.md)Max | 81.41% | SE ±0.21 pp | $0.00331 | 2.58source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 81.412 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.003312 \| USD \| per_test \| Lower is better \| \| `latency` \| 2.575 \| 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: ±0.21 pp sourcePopulation 151 stderr 0.205 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/interpretation_tasks/anthropic~1claude-sonnet-5` |  |  |  |  |  |  |
| 24 | [DeepSeek V4.1 Flash](https://benchize.com/models/deepseek-v4-1-flash-high.md)High | 80.45% | SE ±0.2 pp | $0.00124 | 5.27source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 80.448 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.001236 \| USD \| per_test \| Lower is better \| \| `latency` \| 5.273 \| 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: ±0.2 pp sourcePopulation 151 stderr 0.203 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/interpretation_tasks/deepseek~1deepseek-v4.1-flash` |  |  |  |  |  |  |
| 25 | [Gemini 3.5 Flash-Lite](https://benchize.com/models/gemini-3-5-flash-lite-high.md)High | 80.33% | SE ±0.34 pp | $0.00118 | 1.49source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 80.332 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.001183 \| USD \| per_test \| Lower is better \| \| `latency` \| 1.492 \| 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: ±0.34 pp sourcePopulation 151 stderr 0.336 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/interpretation_tasks/google~1gemini-3.5-flash-lite` |  |  |  |  |  |  |
| 26 | [MiMo-V2.5-Pro](https://benchize.com/models/mimo-v2-5-pro.md) | 79.5% | SE ±0.22 pp | $0.000979 | 31.88source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 79.497 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.000979 \| USD \| per_test \| Lower is better \| \| `latency` \| 31.877 \| 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: ±0.22 pp sourcePopulation 151 stderr 0.223 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/interpretation_tasks/xiaomi~1mimo-v2.5-pro` |  |  |  |  |  |  |
| 27 | [Gemini 3.1 Flash Lite Preview](https://benchize.com/models/gemini-3-1-flash-lite-preview-high.md)High | 78.29% | SE ±0.34 pp | $0.000309 | 2.73source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 78.293 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.000309 \| USD \| per_test \| Lower is better \| \| `latency` \| 2.731 \| 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: ±0.34 pp sourcePopulation 151 stderr 0.344 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/interpretation_tasks/google~1gemini-3.1-flash-lite-preview` |  |  |  |  |  |  |
| 28 | [Nemotron 3 Ultra](https://benchize.com/models/nemotron-3-ultra-550b-a55b.md) | 78.27% | SE ±0.21 pp | — | 6.23source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 78.266 \| percent \| reported \| Higher is better \| \| `latency` \| 6.225 \| 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: ±0.21 pp sourcePopulation 151 stderr 0.209 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/interpretation_tasks/nvidia~1nemotron-3-ultra-550b-a55b` |  |  |  |  |  |  |
| 29 | [MiniMax M2.5](https://benchize.com/models/minimax-m2-5.md) | 78.02% | SE ±0.2 pp | $0.00116 | 13.91source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 78.02 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.001159 \| USD \| per_test \| Lower is better \| \| `latency` \| 13.91 \| 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: ±0.2 pp sourcePopulation 151 stderr 0.204 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/interpretation_tasks/minimax~1MiniMax-M2.5` |  |  |  |  |  |  |
| 30 | [GPT-4o (2024-08-06)](https://benchize.com/models/gpt-4o-2024-08-06-high.md)High | 77.97% | SE ±0.2 pp | — | 0.37source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 77.968 \| percent \| reported \| Higher is better \| \| `latency` \| 0.367 \| 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: ±0.2 pp sourcePopulation 151 stderr 0.199 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/interpretation_tasks/openai~1gpt-4o-2024-08-06` |  |  |  |  |  |  |
| 31 | [GPT-4o (2024-11-20)](https://benchize.com/models/gpt-4o-2024-11-20-high.md)High | 77.91% | SE ±0.21 pp | — | 0.33source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 77.91 \| percent \| reported \| Higher is better \| \| `latency` \| 0.325 \| 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: ±0.21 pp sourcePopulation 151 stderr 0.209 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/interpretation_tasks/openai~1gpt-4o-2024-11-20` |  |  |  |  |  |  |
| 32 | [DeepSeek V4 Pro 0813](https://benchize.com/models/deepseek-v4-pro-0813-max.md)Max | 77.55% | SE ±0.24 pp | $0.00345 | 15.49source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 77.552 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.003453 \| USD \| per_test \| Lower is better \| \| `latency` \| 15.491 \| 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: ±0.24 pp sourcePopulation 151 stderr 0.237 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/interpretation_tasks/deepseek~1deepseek-v4-pro-0813` |  |  |  |  |  |  |
| 33 | [Command A](https://benchize.com/models/command-a.md) | 77.41% | SE ±0.21 pp | — | 0.7source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 77.414 \| percent \| reported \| Higher is better \| \| `latency` \| 0.696 \| 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: ±0.21 pp sourcePopulation 151 stderr 0.209 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/interpretation_tasks/cohere~1command-a-03-2025` |  |  |  |  |  |  |
| 34 | [Gemini 2.5 Flash Lite Preview 09-2025](https://benchize.com/models/gemini-2-5-flash-lite-preview-09-2025.md) | 76.62% | SE ±0.19 pp | — | 0.37source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 76.624 \| percent \| reported \| Higher is better \| \| `latency` \| 0.369 \| 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: ±0.19 pp sourcePopulation 151 stderr 0.192 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/interpretation_tasks/google~1gemini-2.5-flash-lite-preview-09-2025` |  |  |  |  |  |  |
| 35 | [Mistral Large 3 2512](https://benchize.com/models/mistral-large-2512.md) | 76.59% | SE ±0.21 pp | $0.000401 | 0.89source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 76.594 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.000401 \| USD \| per_test \| Lower is better \| \| `latency` \| 0.889 \| 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: ±0.21 pp sourcePopulation 151 stderr 0.207 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/interpretation_tasks/mistralai~1mistral-large-2512` |  |  |  |  |  |  |
| 36 | [GPT-4 Turbo](https://benchize.com/models/gpt-4-turbo-high.md)High | 76.5% | SE ±0.23 pp | — | 0.94source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 76.501 \| percent \| reported \| Higher is better \| \| `latency` \| 0.939 \| 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: ±0.23 pp sourcePopulation 151 stderr 0.233 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/interpretation_tasks/openai~1gpt-4-turbo` |  |  |  |  |  |  |
| 37 | [Laguna M 1](https://benchize.com/models/laguna-m-1.md) | 75.6% | SE ±0.23 pp | — | 37.17source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 75.595 \| percent \| reported \| Higher is better \| \| `latency` \| 37.165 \| 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: ±0.23 pp sourcePopulation 151 stderr 0.229 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/interpretation_tasks/poolside~1laguna-m.1` |  |  |  |  |  |  |
| 38 | [DeepSeek V4 Pro](https://benchize.com/models/deepseek-v4-pro-max.md)Max | 75.17% | SE ±0.23 pp | $0.00551 | 242.07source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 75.169 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.005512 \| USD \| per_test \| Lower is better \| \| `latency` \| 242.068 \| 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: ±0.23 pp sourcePopulation 151 stderr 0.226 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/interpretation_tasks/deepseek~1deepseek-v4-pro` |  |  |  |  |  |  |
| 39 | [MiMo-V2.5](https://benchize.com/models/mimo-v2-5.md) | 74.35% | SE ±0.24 pp | $0.000317 | 12.89source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 74.351 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.000317 \| USD \| per_test \| Lower is better \| \| `latency` \| 12.892 \| 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: ±0.24 pp sourcePopulation 151 stderr 0.235 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/interpretation_tasks/xiaomi~1mimo-v2.5` |  |  |  |  |  |  |
| 40 | [MiniMax M2.1](https://benchize.com/models/minimax-m2-1.md) | 73.59% | SE ±0.26 pp | $0.000753 | 9.03source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 73.592 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.000753 \| USD \| per_test \| Lower is better \| \| `latency` \| 9.034 \| 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: ±0.26 pp sourcePopulation 151 stderr 0.263 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/interpretation_tasks/minimax~1MiniMax-M2.1` |  |  |  |  |  |  |
| 41 | [DeepSeek V4 Flash 0731](https://benchize.com/models/deepseek-v4-flash-0731-high.md)High | 70.99% | SE ±0.24 pp | $0.000592 | 3.94source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 70.993 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.000592 \| USD \| per_test \| Lower is better \| \| `latency` \| 3.943 \| 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: ±0.24 pp sourcePopulation 151 stderr 0.236 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/interpretation_tasks/deepseek~1deepseek-v4-flash-0731` |  |  |  |  |  |  |
| 42 | [Laguna Xs 2](https://benchize.com/models/laguna-xs-2.md) | 69.95% | SE ±0.27 pp | — | 12.12source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 69.947 \| percent \| reported \| Higher is better \| \| `latency` \| 12.122 \| 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: ±0.27 pp sourcePopulation 151 stderr 0.267 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/interpretation_tasks/poolside~1laguna-xs.2` |  |  |  |  |  |  |
| 43 | [GPT-3.5 Turbo](https://benchize.com/models/gpt-3-5-turbo-high.md)High | 65.52% | SE ±0.21 pp | — | 0.46source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 65.515 \| percent \| reported \| Higher is better \| \| `latency` \| 0.46 \| 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: ±0.21 pp sourcePopulation 151 stderr 0.21 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/interpretation_tasks/openai~1gpt-3.5-turbo` |  |  |  |  |  |  |
| 44 | [Command A Plus](https://benchize.com/models/command-a-plus.md) | 54.45% | SE ±0.29 pp | $0.01 | 7.84source duration |  |
| ### Original source values \| Metric \| Value \| Unit \| Scope \| Direction \| \| --- \| --- \| --- \| --- \| --- \| \| `accuracy` \| 54.446 \| percent \| reported \| Higher is better \| \| `cost_per_test` \| 0.011865 \| USD \| per_test \| Lower is better \| \| `latency` \| 7.838 \| 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: ±0.29 pp sourcePopulation 151 stderr 0.288 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/interpretation_tasks/cohere~1command-a-plus-05-2026` |  |  |  |  |  |  |

---
Cite as: LegalBench / interpretation tasks — Benchize, https://benchize.com/benchmarks/vals-ai%3Alegal_bench%3Ainterpretation_tasks (retrieved 2026-10-21)