Benchmark Almanac

MedQA-Followup

GPT-5 mini 0% leads since October 2025
Not in model cards

Models answer medical licensing questions and then respond to follow-ups that challenge their answers. Researchers compare accuracy before and after each intervention.

In the table checked November 19, 2026, GPT-5 scored 95.0% after conflicting retrieval context, followed by Gemini 3 Pro at 89.0%. GPT-4o scored 9.4%.

Made by
Dynamo AI and collaborators
Released
October 2025
Kind
Medicine
Level
Expert
Scored
Checked automatically
Answers
Multiple choice
Tasks
All public
Measures
Capability
Language
English
In and out
Text → Text
Interaction
Multi-turn
Refresh
One fixed set
Run by
Anyone
Based on
MedQA
Occupations
Healthcare practitioners and technical
Strongest model tested
Gemini 3 Pro #56 on the capability index
Last model added
November 2025
Scores refreshed
November 20th, 2026
15
Entries on its own leaderboard, 12 models we track

GPT-5 mini leads with 0% among models released by November 2025 GPT-5 mini leads with 0% among models released by November 2025

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 November 2025, 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 mini low 0%
  2. 2 GPT-5 -0.86%
  3. 3 gpt-oss-120b -1.71%
  4. 4 Gemini 3 Pro medium -6.67%
  5. 5 gpt-oss-20b -8.86%
  6. 6 Gemini 2.5 Pro Jun 2025 -18.57%
  7. 7 Claude Opus 4.5 -33.33%
  8. 8 Gemini 2.5 Flash Sep 2025 -36.10%
  9. 9 Grok 4 Fast -36.76%
  10. 10 MedGemma 4B -42.29%
  11. 11 MedGemma 27B -54.00%
  12. 12 Claude Sonnet 4.5 medium -56.95%
  13. 13 Qwen2.5 72B Turbo -68.86%
  14. 14 GPT-4o Mar 2025 -81.43%
  15. 15 Claude Sonnet 4 -81.81%

Source: Official MedQA-Followup leaderboard

#EntryScoreSettingDateSource
1 GPT-5 mini 0% low Aug 2025 Official leaderboard
2 GPT-5 -0.86% Aug 2025 Official leaderboard
3 gpt-oss-120b -1.71% Aug 2025 Official leaderboard
4 Gemini 3 Pro -6.67% medium Nov 2025 Official leaderboard
5 gpt-oss-20b -8.86% Aug 2025 Official leaderboard
6 Gemini 2.5 Pro Jun 2025 -18.57% Jun 2025 Official leaderboard
7 Claude Opus 4.5 -33.33% Nov 2025 Official leaderboard
8 Gemini 2.5 Flash Sep 2025 -36.10% Sep 2025 Official leaderboard
9 Grok 4 Fast -36.76% Sep 2025 Official leaderboard
10 MedGemma 4B -42.29% Official leaderboard
11 MedGemma 27B -54.00% Official leaderboard
12 Claude Sonnet 4.5 -56.95% medium Sep 2025 Official leaderboard
13 Qwen2.5 72B Turbo -68.86% Official leaderboard
14 GPT-4o Mar 2025 -81.43% Mar 2025 Official leaderboard
15 Claude Sonnet 4 -81.81% May 2025 Official leaderboard
#ModelScoreSettingDateSource
1 GPT-5 mini 0% low Aug 2025 Official leaderboard
2 GPT-5 -0.86% Aug 2025 Official leaderboard
3 gpt-oss-120b -1.71% Aug 2025 Official leaderboard
4 Gemini 3 Pro -6.67% medium Nov 2025 Official leaderboard
5 gpt-oss-20b -8.86% Aug 2025 Official leaderboard
6 Gemini 2.5 Pro Jun 2025 -18.57% Jun 2025 Official leaderboard
7 Claude Opus 4.5 -33.33% Nov 2025 Official leaderboard
8 Gemini 2.5 Flash Sep 2025 -36.10% Sep 2025 Official leaderboard
9 Grok 4 Fast -36.76% Sep 2025 Official leaderboard
10 Claude Sonnet 4.5 -56.95% medium Sep 2025 Official leaderboard
11 GPT-4o Mar 2025 -81.43% Mar 2025 Official leaderboard
12 Claude Sonnet 4 -81.81% May 2025 Official leaderboard

Rankings by part

We read 8 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as Authority Prior, Autograder Prior and Commitment Alignment. gpt-oss-20b leads 3 of them, and 3 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.

Authority Prior

12 models we track have a score here.

  1. 1 gpt-oss-120b ↑2 0.86%
  2. 2 GPT-5 mini ↓1 0%
  3. 3 GPT-5 ↓1 -0.10%
  4. 4 gpt-oss-20b ↑1 -0.86%
  5. 5 Grok 4 Fast ↑4 -1.90%
  6. 6 Gemini 3 Pro ↓2 -2.86%
  7. 7 GPT-4o Mar 2025 ↑4 -4.86%
  8. 8 Gemini 2.5 Pro Jun 2025 ↓2 -7.05%
  9. 9 Claude Opus 4.5 ↓2 -7.62%
  10. 10 Gemini 2.5 Flash Sep 2025 ↓2 -8.86%

Number one over time

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

OpenAI
Own leaderboard Not tested since Nov 2025

Source: Official MedQA-Followup 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%) GPT-5 mini GPT-4o 0% 20% 40% 60% 80% 100% 2025 2026 Near the ceiling (85%) GPT-5 mini

Source: Official MedQA-Followup 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

9 of 10 checks pass, 1 warning. Look first at the Newest score check. The newest came 12 months ago, in November 2025, for Claude Opus 4.5. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 1 warning 3 pass 1 note
  • Pass Independent runs Scores from the official leaderboard (15 entries).
  • Pass Models scored 12 models we track, plus 3 other entries on its leaderboard.
  • Warning Newest score The newest came 12 months ago, in November 2025, for Claude Opus 4.5.
  • Note Frontier models None of the top 10 has a score. The strongest tested is Gemini 3 Pro (#56).
  • Pass Headroom The best is 0%, under 85%, so it still separates models.
Leaderboard import 3 pass
  • Pass Importer We read the board on bmanczak.github.io: 14 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