Benchmark Almanac

PSEBench

Gemini 3.1 Pro 95.3% leads since June 2026
Not in model cards

Models triage patient safety reports under Minnesota rules. They identify reportable events, request missing facts, or abstain when the evidence permits multiple verdicts.

The authors evaluated 15 models on 5,074 cases.

Made by
Scale AI and collaborators
Released
June 2026
Kind
Medicine
Level
Expert
Scored
Checked automatically
Answers
Agent task
Measures
Capability
Subject
Health sciences, Law
Language
English
In and out
Text → Text, Actions
Interaction
Agentic
Environment
A simulator
Refresh
One fixed set
Run by
A host
Occupations
Healthcare practitioners and technical
Strongest model tested
GPT-5.5 #21 on the capability index
Last model added
April 2026
Scores refreshed
November 20th, 2026
15
Entries in its paper, 12 models we track

Gemini 3.1 Pro leads with 95.3% among models released by April 2026 Gemini 3.1 Pro leads with 95.3% among models released by April 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 with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. This view shows only models we track, each at its best.

The newest model with a score here came out in April 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 Gemini 3.1 Pro 95.3%
  2. 2 GPT-5.5 94.6%
  3. 3 Claude Opus 4.7 92.3%
  4. 4 GPT-5 90.2%
  5. 5 Gemini 2.5 Flash Sep 2025 85.8%
  6. 6 Claude Sonnet 4.6 84.5%
  7. 7 DeepSeek-R1 May 2025 83.7%
  8. 8 gpt-oss-120b 79%
  9. 9 Qwen3-235B-A22B 76.7%
  10. 10 GPT-5 nano 71.9%
  11. 11 HuatuoGPT-o1 (70B) 61.1%
  12. 12 Mistral Small 3.2 60.5%
  13. 13 Llama 3.1-8B 52.4%
  14. 14 MedGemma (27B) 51.3%
  15. 15 HuatuoGPT-o1 (8B) 45.5%

Source: The PSEBench paper

#EntryScoreSettingDateSource
1 Gemini 3.1 Pro 95.3% Feb 2026 Official leaderboard
2 GPT-5.5 94.6% Apr 2026 Official leaderboard
3 Claude Opus 4.7 92.3% Apr 2026 Official leaderboard
4 GPT-5 90.2% Aug 2025 Official leaderboard
5 Gemini 2.5 Flash Sep 2025 85.8% Sep 2025 Official leaderboard
6 Claude Sonnet 4.6 84.5% Feb 2026 Official leaderboard
7 DeepSeek-R1 May 2025 83.7% May 2025 Official leaderboard
8 gpt-oss-120b 79% Aug 2025 Official leaderboard
9 Qwen3-235B-A22B 76.7% Apr 2025 Official leaderboard
10 GPT-5 nano 71.9% Aug 2025 Official leaderboard
11 HuatuoGPT-o1 (70B) 61.1% Official leaderboard
12 Mistral Small 3.2 60.5% Jun 2025 Official leaderboard
13 Llama 3.1-8B 52.4% Jul 2024 Official leaderboard
14 MedGemma (27B) 51.3% Official leaderboard
15 HuatuoGPT-o1 (8B) 45.5% Official leaderboard
#ModelScoreSettingDateSource
1 Gemini 3.1 Pro 95.3% Feb 2026 Official leaderboard
2 GPT-5.5 94.6% Apr 2026 Official leaderboard
3 Claude Opus 4.7 92.3% Apr 2026 Official leaderboard
4 GPT-5 90.2% Aug 2025 Official leaderboard
5 Gemini 2.5 Flash Sep 2025 85.8% Sep 2025 Official leaderboard
6 Claude Sonnet 4.6 84.5% Feb 2026 Official leaderboard
7 DeepSeek-R1 May 2025 83.7% May 2025 Official leaderboard
8 gpt-oss-120b 79% Aug 2025 Official leaderboard
9 Qwen3-235B-A22B 76.7% Apr 2025 Official leaderboard
10 GPT-5 nano 71.9% Aug 2025 Official leaderboard
11 Mistral Small 3.2 60.5% Jun 2025 Official leaderboard
12 Llama 3.1-8B 52.4% Jul 2024 Official leaderboard

Rankings by part

We read 10 parts from the benchmark’s paper, each ranking models on one part of the test, such as Clause accuracy, Evidence citation F1 and Boundary hit rate. GPT-5.5 leads 5 of them, and 2 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.

Clause accuracy

12 models we track have a score here.

  1. 1 GPT-5.5 ↑1 98.5%
  2. 2 GPT-5 ↑2 98%
  3. 3 Claude Sonnet 4.6 ↑3 97.9%
  4. 4 Gemini 3.1 Pro ↓3 97.8%
  5. 5 Claude Opus 4.7 ↓2 97.3%
  6. 6 Gemini 2.5 Flash Sep 2025 ↓1 96.7%
  7. 7 DeepSeek-R1 May 2025 95.3%
  8. 8 gpt-oss-120b 94.3%
  9. 9 Qwen3-235B-A22B 93.3%
  10. 10 Mistral Small 3.2 ↑1 89.5%

Number one over time

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

Google DeepMind
Benchmark paper Passed 85% in Jun 2026, by Gemini 3.1 Pro

Source: The PSEBench paper

GPT-5 on the official leaderboard passed 85% in August 2025

Models released before it came out already scored that high. 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% 2025 2026 Near the ceiling (85%) Gemini 3.1 Pro GPT-5 DeepSeek-R1 Qwen3-235B-A22B Llama 3.1-8B 0% 20% 40% 60% 80% 100% 2025 2026 Near the ceiling (85%) Gemini 3.1 Pro DeepSeek-R1

Source: Official PSEBench 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

7 of 8 checks pass, 1 warning. Look first at the Task access check. We don’t know whether the tasks are public. Every benchmark

Catalog record 1 warning 1 pass
  • Warning Task access We don’t know whether the tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 2 pass 3 notes
  • Pass Independent runs Scores from the benchmark’s paper (15 entries).
  • Pass Models scored 12 models we track, plus 3 other entries on its leaderboard.
  • Note Newest score The newest came 7 months ago, in April 2026, for GPT-5.5. That’s expected for a saturated test.
  • Note Frontier models None of the top 10 has a score, which is fine for a retired test. The strongest tested is GPT-5.5 (#21).
  • Note Headroom Saturated. The best is 95.3%, past 85% since August 2025.
Leaderboard import 3 pass
  • Pass Importer We read the board on labs.scale.com: 12 rows for models we track and 3 other entries 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