Benchmark Almanac

JusticeAxis

GPT-5.6 91% tops its leaderboard Gemini 3.8 Flash 84.4% leads the models we track
Not in model cards

Models assess evidence and produce judgments for 256 criminal cases from 18 countries. The evaluator checks charges and outcomes against lawyer-written references.

Released
September 2026
Kind
Law
Level
Expert
Scored
Checked automatically
Answers
Open-ended
Tasks
All public
Measures
Capability
Subject
Law
In and out
Text, Images, Audio → Text
Interaction
Single turn
Refresh
One fixed set
Run by
Anyone
Occupations
Legal
Strongest model tested
Gemini 3.8 Flash #24 on the capability index
Last model added
September 2026
Scores refreshed
November 20th, 2026
13
Entries in its paper, 6 models we track

GPT-5.6 tops the leaderboard with 91% Gemini 3.8 Flash leads with 84.4% among models released by September 2026

Best score per entry from the benchmark's paper, read on November 20th, 2026. A paper’s table doesn’t add newer models. Where a model ran on its own with several settings, we keep its best. Agents other companies built around a model get a pale shade of that model's lab color. Gray marks blends and entries not in our model list, such as fine-tunes and older models. None of these count toward leaders. The best model we track is Gemini 3.8 Flash, at 84.4%. Agents built around a model, such as JusticeAgent + Qwen 3.8 27B, are their own entries. We don't count them as the model's score. This view shows only models we track, each at its best.

The newest model with a score here came out in September 2026, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

  1. 1 GPT-5.6 91%
  2. 2 Gemini 3.8 Flash 84.4%
  3. 3 Grok 4.6 83.2%
  4. 4 JusticeAgent + Qwen 3.8 27B 80.9%
  5. 5 Claude Sonnet 5 78.5%
  6. 6 Element agents + Qwen 3.8 27B 76.8%
  7. 7 Multimodal agents + Qwen 3.8 27B 74.8%
  8. 8 Agentic legal search + Qwen 3.8 27B 71.5%
  9. 9 Courtroom simulation + Qwen 3.8 27B 70.1%
  10. 10 Syllogistic retrieval + Qwen 3.8 27B 68.8%
  11. 11 DeepSeek-V4-Flash 58.4%
  12. 12 Qwen 3.8 27B Syllogism prompting 57.4%
  13. 13 Gemma 4 31B IT 32%

Source: The JusticeAxis paper

#EntryScoreSettingDateSource
1 GPT-5.6 91% Official leaderboard
2 Gemini 3.8 Flash 84.4% Sep 2026 Official leaderboard
3 Grok 4.6 83.2% Aug 2026 Official leaderboard
4 JusticeAgent + Qwen 3.8 27B 80.9% Aug 2026 Official leaderboard
5 Claude Sonnet 5 78.5% Jun 2026 Official leaderboard
6 Element agents + Qwen 3.8 27B 76.8% Aug 2026 Official leaderboard
7 Multimodal agents + Qwen 3.8 27B 74.8% Aug 2026 Official leaderboard
8 Agentic legal search + Qwen 3.8 27B 71.5% Aug 2026 Official leaderboard
9 Courtroom simulation + Qwen 3.8 27B 70.1% Aug 2026 Official leaderboard
10 Syllogistic retrieval + Qwen 3.8 27B 68.8% Aug 2026 Official leaderboard
11 DeepSeek-V4-Flash 58.4% Apr 2026 Official leaderboard
12 Qwen 3.8 27B 57.4% Syllogism prompting Aug 2026 Official leaderboard
13 Gemma 4 31B IT 32% Apr 2026 Official leaderboard
#ModelScoreSettingDateSource
1 Gemini 3.8 Flash 84.4% Sep 2026 Official leaderboard
2 Grok 4.6 83.2% Aug 2026 Official leaderboard
3 Claude Sonnet 5 78.5% Jun 2026 Official leaderboard
4 DeepSeek-V4-Flash 58.4% Apr 2026 Official leaderboard
5 Qwen 3.8 27B 57.4% Syllogism prompting Aug 2026 Official leaderboard
6 Gemma 4 31B IT 32% Apr 2026 Official leaderboard

2 models have held number one in the benchmark’s paper since its release in September 2026

Lab claims count from the day the card came out, independent runs from the model’s release.

xAI Google DeepMind
Benchmark paper Not tested since Sep 2026

Source: The JusticeAxis paper

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

Plot with other benchmarks
In the benchmark’s paper Best so far
0% 20% 40% 60% 80% 100% 2026 Near the ceiling (85%) Gemini 3.8 Flash Claude Sonnet 5 DeepSeek-V4-Flash Gemma 4 31B IT 0% 20% 40% 60% 80% 100% 2026 Near the ceiling (85%) Gemini 3.8 Flash DeepSeek-V4-Flash

Source: Official JusticeAxis leaderboard

No lab model card reports it yet

We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.

More in Law, finance and medicine

All 637 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%
  • 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. o1-preview 92.7%
  • Harvey LAB-AA v1.1 In use Artificial Analysis's run of Harvey's Legal Agent Benchmark on Harvey's newer private set of 120 tasks across 24 practice areas. A deliverable with a material hallucination scores zero, and the headline is the share of judges who find every rubric criterion passed. GPT-6.2 Astra 15%

Data audit

All 10 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 4 pass 1 note
  • Pass Independent runs Scores from the benchmark’s paper (13 entries).
  • Pass Models scored 6 models we track, plus 7 other entries on its leaderboard.
  • Pass Newest score The newest is from September 2026, for Gemini 3.8 Flash.
  • Note Frontier models None of the top 10 has a score. The strongest tested is Gemini 3.8 Flash (#24).
  • Pass Headroom The best is 84.4%, under 85%, so it still separates models.
Leaderboard import 3 pass
  • Pass Importer We read the board on arxiv.org: 13 rows for models we track and 1 other entry at the last read.
  • Pass Last read November 20th, 2026.
  • Pass Read worked The last read worked.
Consistency 1 pass 1 note
  • Note Lab claims No model has both a lab number and an independent run.
  • Pass Known issues No published audit found problems.

How we check