Benchmark Almanac

KlinikeBench

Claude Opus 5 29.4% leads since September 2026
Not in model cards

Agents interview simulated patients, use clinical tools, and record diagnoses across 333 cases. A case passes only when the agent satisfies all four clinical criteria.

The October 6, 2026 paper compares 31 models. Claude Opus 5 led at 29.4% strict pass@1 despite reaching 90.7% diagnostic accuracy.

Made by
Xueting Fang and colleagues
Released
September 2026
Kind
Medicine
Level
Expert
Scored
Graded by a model
Answers
Agent task
Tasks
All public
Measures
Capability
Language
English
In and out
Text → Text, Actions
Interaction
Agentic
Items from
Generated
Refresh
One fixed set
Run by
Anyone
Occupations
Healthcare practitioners and technical
Strongest model tested
GPT-6 Astra #4 on the capability index
Last model added
September 2026
Scores refreshed
November 20th, 2026
31
Entries on its own leaderboard, 29 models we track

Claude Opus 5 leads with 29.4% among models released by September 2026 Claude Opus 5 leads with 29.4% among models released by September 2026

Best score per entry from the benchmark's own leaderboard, read on November 20th, 2026. 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 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 Claude Opus 5 29.4%
  2. 2 Claude Opus 5.5 28.8%
  3. 3 GPT-6 Astra 24.3%
  4. 4 GPT-5.6 Sol 23.7%
  5. 5 GPT-5.4 23.4%
  6. 6 Kimi K2.6 23.4%
  7. 7 Grok 4.7 22.2%
  8. 8 Grok 4.6 21.9%
  9. 9 Grok 4.3 21.3%
  10. 10 Gemini 3.8 Flash 18.9%
  11. 11 GPT-5.5 16.8%
  12. 12 Claude Opus 4.8 16.8%
  13. 13 GPT-5.4 Mini 15.9%
  14. 14 Gemini 3.5 Flash 15.6%
  15. 15 GPT-5.1 15.3%
  16. 16 Claude Sonnet 5 13.8%
  17. 17 Gemini 3.7 Flash 10.8%
  18. 18 Claude Sonnet 4.5 9.90%
  19. 19 Claude Sonnet 4.6 8.40%
  20. 20 GPT-5.6 Terra 8.40%
  21. 21 DeepSeek-V4-Pro 8.10%
  22. 22 GPT-4.1 7.50%
  23. 23 GPT-5.6 Luna 7.50%
  24. 24 GLM-5.1 4.80%
  25. 25 GPT-4o (snapshot unspecified) 3.90%
  26. 26 DeepSeek-V4-Flash 3.90%
  27. 27 GPT-5.3 Codex 3.60%
  28. 28 Claude Haiku 4.5 2.70%
  29. 29 GPT-4.1 mini 2.10%
  30. 30 GPT-4.1 nano 0.90%
  31. 31 GPT-4o-mini (snapshot unspecified) 0.60%

Source: Official KlinikeBench leaderboard

#EntryScoreSettingDateSource
1 Claude Opus 5 29.4% Terminus 2 Jul 2026 Official leaderboard
2 Claude Opus 5.5 28.8% Terminus 2 Sep 2026 Official leaderboard
3 GPT-6 Astra 24.3% Terminus 2 Sep 2026 Official leaderboard
4 GPT-5.6 Sol 23.7% Terminus 2 Jul 2026 Official leaderboard
5 GPT-5.4 23.4% Terminus 2 Mar 2026 Official leaderboard
6 Kimi K2.6 23.4% Terminus 2 Apr 2026 Official leaderboard
7 Grok 4.7 22.2% Terminus 2 Sep 2026 Official leaderboard
8 Grok 4.6 21.9% Terminus 2 Aug 2026 Official leaderboard
9 Grok 4.3 21.3% Terminus 2 Apr 2026 Official leaderboard
10 Gemini 3.8 Flash 18.9% Terminus 2 Sep 2026 Official leaderboard
11 GPT-5.5 16.8% Terminus 2 Apr 2026 Official leaderboard
12 Claude Opus 4.8 16.8% Terminus 2 May 2026 Official leaderboard
13 GPT-5.4 Mini 15.9% Terminus 2 Mar 2026 Official leaderboard
14 Gemini 3.5 Flash 15.6% Terminus 2 May 2026 Official leaderboard
15 GPT-5.1 15.3% Terminus 2 Nov 2025 Official leaderboard
16 Claude Sonnet 5 13.8% Terminus 2 Jun 2026 Official leaderboard
17 Gemini 3.7 Flash 10.8% Terminus 2 Aug 2026 Official leaderboard
18 Claude Sonnet 4.5 9.90% Terminus 2 Sep 2025 Official leaderboard
19 Claude Sonnet 4.6 8.40% Terminus 2 Feb 2026 Official leaderboard
20 GPT-5.6 Terra 8.40% Terminus 2 Jul 2026 Official leaderboard
21 DeepSeek-V4-Pro 8.10% Terminus 2 Apr 2026 Official leaderboard
22 GPT-4.1 7.50% Terminus 2 Apr 2025 Official leaderboard
23 GPT-5.6 Luna 7.50% Terminus 2 Jul 2026 Official leaderboard
24 GLM-5.1 4.80% Terminus 2 Apr 2026 Official leaderboard
25 GPT-4o (snapshot unspecified) 3.90% Terminus 2 Official leaderboard
26 DeepSeek-V4-Flash 3.90% Terminus 2 Apr 2026 Official leaderboard
27 GPT-5.3 Codex 3.60% Terminus 2 Feb 2026 Official leaderboard
28 Claude Haiku 4.5 2.70% Terminus 2 Oct 2025 Official leaderboard
29 GPT-4.1 mini 2.10% Terminus 2 Apr 2025 Official leaderboard
30 GPT-4.1 nano 0.90% Terminus 2 Apr 2025 Official leaderboard
31 GPT-4o-mini (snapshot unspecified) 0.60% Terminus 2 Official leaderboard
#ModelScoreSettingDateSource
1 Claude Opus 5 29.4% Terminus 2 Jul 2026 Official leaderboard
2 Claude Opus 5.5 28.8% Terminus 2 Sep 2026 Official leaderboard
3 GPT-6 Astra 24.3% Terminus 2 Sep 2026 Official leaderboard
4 GPT-5.6 Sol 23.7% Terminus 2 Jul 2026 Official leaderboard
5 GPT-5.4 23.4% Terminus 2 Mar 2026 Official leaderboard
6 Kimi K2.6 23.4% Terminus 2 Apr 2026 Official leaderboard
7 Grok 4.7 22.2% Terminus 2 Sep 2026 Official leaderboard
8 Grok 4.6 21.9% Terminus 2 Aug 2026 Official leaderboard
9 Grok 4.3 21.3% Terminus 2 Apr 2026 Official leaderboard
10 Gemini 3.8 Flash 18.9% Terminus 2 Sep 2026 Official leaderboard
11 GPT-5.5 16.8% Terminus 2 Apr 2026 Official leaderboard
12 Claude Opus 4.8 16.8% Terminus 2 May 2026 Official leaderboard
13 GPT-5.4 Mini 15.9% Terminus 2 Mar 2026 Official leaderboard
14 Gemini 3.5 Flash 15.6% Terminus 2 May 2026 Official leaderboard
15 GPT-5.1 15.3% Terminus 2 Nov 2025 Official leaderboard
16 Claude Sonnet 5 13.8% Terminus 2 Jun 2026 Official leaderboard
17 Gemini 3.7 Flash 10.8% Terminus 2 Aug 2026 Official leaderboard
18 Claude Sonnet 4.5 9.90% Terminus 2 Sep 2025 Official leaderboard
19 Claude Sonnet 4.6 8.40% Terminus 2 Feb 2026 Official leaderboard
20 GPT-5.6 Terra 8.40% Terminus 2 Jul 2026 Official leaderboard
21 DeepSeek-V4-Pro 8.10% Terminus 2 Apr 2026 Official leaderboard
22 GPT-4.1 7.50% Terminus 2 Apr 2025 Official leaderboard
23 GPT-5.6 Luna 7.50% Terminus 2 Jul 2026 Official leaderboard
24 GLM-5.1 4.80% Terminus 2 Apr 2026 Official leaderboard
25 DeepSeek-V4-Flash 3.90% Terminus 2 Apr 2026 Official leaderboard
26 GPT-5.3 Codex 3.60% Terminus 2 Feb 2026 Official leaderboard
27 Claude Haiku 4.5 2.70% Terminus 2 Oct 2025 Official leaderboard
28 GPT-4.1 mini 2.10% Terminus 2 Apr 2025 Official leaderboard
29 GPT-4.1 nano 0.90% Terminus 2 Apr 2025 Official leaderboard

Rankings by part

We read 3 parts from the benchmark’s own leaderboard: Diagnosis, Required tools and Must-ask. Each ranks models on one part of the test. Claude Opus 5 leads 1 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.

Diagnosis

29 models we track have a score here.

  1. 1 Claude Opus 5 90.7%
  2. 2 Claude Opus 5.5 90.7%
  3. 3 Claude Opus 4.8 ↑9 89.2%
  4. 4 Claude Sonnet 5 ↑12 87.4%
  5. 5 GPT-6 Astra ↓2 87.4%
  6. 6 GPT-5.5 ↑5 86.2%
  7. 7 GPT-5.6 Sol ↓3 86.2%
  8. 8 GPT-5.6 Terra ↑12 84.7%
  9. 9 GPT-5.3 Codex ↑17 84.4%
  10. 10 Claude Sonnet 4.6 ↑9 83.8%

Number one over time

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

Anthropic
Own leaderboard Not tested since Sep 2026

Source: Official KlinikeBench leaderboard

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
On the benchmark’s own leaderboard Best so far
0% 20% 40% 60% 80% 100% 2025 2026 Near the ceiling (85%) Claude Opus 5 GPT-5.6 Sol GPT-5.4 GPT-5.1 Claude Sonnet 4.5 GPT-4.1 0% 20% 40% 60% 80% 100% 2025 2026 Near the ceiling (85%) Claude Opus 5 GPT-5.1

Source: Official KlinikeBench 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 official leaderboard (31 entries).
  • Pass Models scored 29 models we track, plus 2 other entries on its leaderboard.
  • Pass Newest score The newest is from September 2026, for Claude Opus 5.5.
  • Note Frontier models Only 2 of the top 10 have a score. The strongest tested is GPT-6 Astra (#4).
  • Pass Headroom The best is 29.4%, under 85%, so it still separates models.
Leaderboard import 3 pass
  • Pass Importer We read the board on zehui127.github.io: 29 rows for models we track and 2 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