Legora BAR
Models work private legal cases end to end inside Legora's agent harness, with the matter's documents and templates. An LLM judge scores outputs against rubrics written by lawyers.
Cases span 30 practice areas and seven task families. Results are shown relative to the average of the models run, 17 models in the 2026-10 edition. Legora reruns it as models and harness change.
- Made by
- Legora
- Released
- September 2026
- Kind
- Law
- Level
- Expert
- Scored
- Graded by a model
- Answers
- Agent task
- Tasks
- Kept private
- Strongest model tested
- GPT-6 Astra #3 on the capability index
- Last model added
- October 2026
- Scores refreshed
- November 12th, 2026
Claude Opus 5.5 leads with 106.0
Best score per model from the benchmark's own leaderboard, read on November 12th, 2026. Where a model ran with several settings or harnesses, we keep its best.
- 106.0 Claude Opus 5.5
- 105.0
Muse Wonder - 104.0 Claude Fable 5.1
- 103.0
Gemini 3.8 Flash - 103.0
Mistral Large 4 - 102.0 Kimi K3
- 102.0 GLM-5.3
- 102.0 Grok 4.7
- 102.0 Claude Sonnet 5.5
- 102.0 GPT-6.1 Sol
- 101.0 Muse Spark 1.3
- 100.0
DeepSeek V4.1 Flash - 99.0 Gemini 4 Argon
- 96.0 GPT-6 Astra
- 94.0 GLM-5.3-Flash
- 93.0 DeepSeek-V4-Pro
- 88.0 GPT-6 Luna
- 1 Claude Opus 5.5 2026-10 106.0
- 2 Muse Wonder 2026-10 105.0
- 3 Claude Fable 5.1 2026-10 104.0
- 4 Gemini 3.8 Flash 2026-10 103.0
- 5 Mistral Large 4 2026-10 103.0
- 6 Kimi K3 2026-10 102.0
- 7 GLM-5.3 2026-10 102.0
- 8 Grok 4.7 2026-10 102.0
- 9 Claude Sonnet 5.5 2026-10 102.0
- 10 GPT-6.1 Sol 2026-10 102.0
- 11 Muse Spark 1.3 2026-10 101.0
- 12 DeepSeek V4.1 Flash 2026-10 100.0
- 13 Gemini 4 Argon 2026-10 99.0
- 14 GPT-6 Astra 2026-10 96.0
- 15 GLM-5.3-Flash 2026-10 94.0
- 16 DeepSeek-V4-Pro 2026-10 93.0
- 17 GPT-6 Luna 2026-10 88.0
Source: Official Legora BAR leaderboard
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Claude Opus 5.5 | 106.0 | 2026-10 | Sep 2026 | Official leaderboard |
| 2 | Muse Wonder | 105.0 | 2026-10 | Oct 2026 | Official leaderboard |
| 3 | Claude Fable 5.1 | 104.0 | 2026-10 | Sep 2026 | Official leaderboard |
| 4 | Gemini 3.8 Flash | 103.0 | 2026-10 | Sep 2026 | Official leaderboard |
| 5 | Mistral Large 4 | 103.0 | 2026-10 | Oct 2026 | Official leaderboard |
| 6 | Kimi K3 | 102.0 | 2026-10 | Jul 2026 | Official leaderboard |
| 7 | GLM-5.3 | 102.0 | 2026-10 | Aug 2026 | Official leaderboard |
| 8 | Grok 4.7 | 102.0 | 2026-10 | Sep 2026 | Official leaderboard |
| 9 | Claude Sonnet 5.5 | 102.0 | 2026-10 | Sep 2026 | Official leaderboard |
| 10 | GPT-6.1 Sol | 102.0 | 2026-10 | Sep 2026 | Official leaderboard |
| 11 | Muse Spark 1.3 | 101.0 | 2026-10 | Sep 2026 | Official leaderboard |
| 12 | DeepSeek V4.1 Flash | 100.0 | 2026-10 | Sep 2026 | Official leaderboard |
| 13 | Gemini 4 Argon | 99.0 | 2026-10 | Sep 2026 | Official leaderboard |
| 14 | GPT-6 Astra | 96.0 | 2026-10 | Sep 2026 | Official leaderboard |
| 15 | GLM-5.3-Flash | 94.0 | 2026-10 | Aug 2026 | Official leaderboard |
| 16 | DeepSeek-V4-Pro | 93.0 | 2026-10 | Apr 2026 | Official leaderboard |
| 17 | GPT-6 Luna | 88.0 | 2026-10 | Sep 2026 | Official leaderboard |
Quality vs. median time per case
Both axes are relative indexes where 100 is the average of the models run; lower time is faster. The line joins the runs that no other run beats on both measures. We redrew this chart from the numbers on the official leaderboard.
| Model | Setting | Quality vs average | Median time per case vs average |
|---|---|---|---|
| Claude Opus 5.5 | 2026-10 | 106 index | 123 index |
| Muse Wonder | 2026-10 | 105 index | 32 index |
| Claude Fable 5.1 | 2026-10 | 104 index | 79 index |
| Gemini 3.8 Flash | 2026-10 | 103 index | 83 index |
| Mistral Large 4 | 2026-10 | 103 index | 330 index |
| GPT-6.1 Sol | 2026-10 | 102 index | 78 index |
| Kimi K3 | 2026-10 | 102 index | 82 index |
| GLM-5.3 | 2026-10 | 102 index | 83 index |
| Grok 4.7 | 2026-10 | 102 index | 92 index |
| Claude Sonnet 5.5 | 2026-10 | 102 index | 164 index |
| Muse Spark 1.3 | 2026-10 | 101 index | 34 index |
| DeepSeek V4.1 Flash | 2026-10 | 100 index | 48 index |
| Gemini 4 Argon | 2026-10 | 99 index | 121 index |
| GPT-6 Astra | 2026-10 | 96 index | 100 index |
| GLM-5.3-Flash | 2026-10 | 94 index | 93 index |
| DeepSeek-V4-Pro | 2026-10 | 93 index | 56 index |
| GPT-6 Luna | 2026-10 | 88 index | 101 index |
2 models have held number one on the official leaderboard since its release in September 2026
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official Legora BAR leaderboard
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official Legora BAR leaderboard
No lab model card reports it yet
We read the headline results of 522 launch posts from 52 labs, and none of them list this benchmark.
More in Law, finance and medicine
- MedXpertQA-MM In use Models answer expert-level medical exam questions that combine images with clinical records and test results. Scoring is accuracy.
Qwen 3.8 Max 80.4% - HealthBench Professional In use 525 physician-authored tasks from real clinician chats (care consult, documentation, research) with physician-adjudicated rubrics. Claude Sonnet 5.5 69.2%
- HealthBench In use 5,000 realistic health conversations graded against physician-written rubrics. A length-adjusted variant penalizes verbosity. GPT-5 67.2%
- FinSearchComp In use Agents search the web to fetch live financial data, look up history, and run multi-step investigations, across 635 expert-written questions. Scored by answer accuracy.
Ling-3.1-flash 81.7% - KBL In use Models answer Korean legal knowledge and reasoning tasks and Korean bar exam questions. It has 7 knowledge tasks, 4 reasoning tasks and 2,510 bar exam items.
Solar Pro 4 77.5% - KorMedMCQA In use Models answer 7,469 multiple-choice questions from Korean licensing exams for doctors, nurses, pharmacists and dentists, held from 2012 to 2024. Solar Pro 4 93.2%