Benchmark Almanac

KorMedMCQA-V

Gemini 3 Pro 96.9% leads since February 2026
Not in model cards

Models answer Korean medical licensing questions using clinical images. The full test contains 1,534 questions and 2,043 images from 2012 through 2023.

The authors publish accuracy for 51 vision language models. The shorter 2022 and 2023 test split contains 304 questions.

Released
February 2026
Kind
Medicine
Level
Expert
Scored
Checked automatically
Answers
Multiple choice
Tasks
All public
Measures
Capability
Language
Korean
In and out
Text, Images → Text
Interaction
Single turn
Items from
Exams and contests
Refresh
One fixed set
Run by
Anyone
Based on
KorMedMCQA
Occupations
Healthcare practitioners and technical
Strongest model tested
GPT-5.2 #51 on the capability index
Last model added
December 2025
Scores refreshed
November 20th, 2026
51
Entries on its own leaderboard, 33 models we track

2 models share the top score of 96.9% among models released by December 2025 2 models share the top score of 96.9% among models released by December 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 December 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 Gemini 3 Pro Reasoning 96.9%
  2. 2 Gemini 3 Flash Reasoning 96.9%
  3. 3 GPT-5 Reasoning 93.9%
  4. 4 GPT-5.2 Reasoning 93.9%
  5. 5 GPT-5 mini Reasoning 90.1%
  6. 6 Qwen3-VL-32B-Thinking Reasoning 83.7%
  7. 7 Qwen3-VL-30B-A3B-Thinking Reasoning 80.4%
  8. 8 GLM-4.6V Reasoning 78.7%
  9. 9 Qwen3-VL-32B-Instruct 76.5%
  10. 10 Qwen3-VL-8B-Thinking Reasoning 74.2%
  11. 11 Ministral 3 14B Reasoning 72.7%
  12. 12 Qwen3-VL-30B-A3B-Instruct 71.7%
  13. 13 InternVL3.5-38B 67.2%
  14. 14 InternVL3.5-30B-A3B 66.4%
  15. 15 Lingshu-32B 66.1%
  16. 16 Qwen3-VL-4B-Thinking Reasoning 65.9%
  17. 17 GLM-4.6V-Flash Reasoning 65.5%
  18. 18 Ministral-3-8B-Reasoning-2512 Reasoning 65.1%
  19. 19 Hulu-Med-32B 64.7%
  20. 20 Qwen2.5-VL-32B 63%
  21. 21 Gemma 3 27B 57.3%
  22. 22 MedGemma-27B-IT 56.8%
  23. 23 Ministral-3-14B-Instruct-2512 56.5%
  24. 24 Hulu-Med-14B 55.1%
  25. 25 Gemma 3 12B 54.3%
  26. 26 Qwen3-VL-8B-Instruct 53.8%
  27. 27 Ministral 3 8B Instruct 53.8%
  28. 28 InternVL3.5-14B 53.6%
  29. 29 InternVL3.5-8B 47.4%
  30. 30 Lingshu-7B 46.6%
  31. 31 Qwen2.5-VL-7B 46.6%
  32. 32 Hulu-Med-4B 46.1%
  33. 33 Qwen3-VL-4B-Instruct 45.7%
  34. 34 Ministral 3 3B Reasoning Reasoning 45.2%
  35. 35 MedGemma-1.5-4B-IT 44.4%
  36. 36 Hulu-Med-7B 43.9%
  37. 37 InternVL3.5-4B 43.3%
  38. 38 VARCO-VISION-2.0-14B 43.2%
  39. 39 A.X-4.0-VL-Light 41.8%
  40. 40 Ministral 3 3B Instruct 41.8%
  41. 41 Qwen3-VL-2B-Thinking Reasoning 40.4%
  42. 42 Kimi-VL-A3B-Thinking Reasoning 37.6%
  43. 43 MedGemma-4B-IT 34.4%
  44. 44 Gemma 3 4B 34.3%
  45. 45 Qwen3-VL-2B-Instruct 31.3%
  46. 46 Kimi-VL-A3B-Instruct 30.8%
  47. 47 Kanana-1.5-V-3B-Instruct 30.7%
  48. 48 InternVL3.5-2B 29.8%
  49. 49 HyperCLOVAX-SEED-Vision-Instruct-3B 26.3%
  50. 50 VARCO-VISION-2.0-1.7B 24.4%
  51. 51 InternVL3.5-1B 22.4%

Source: Official KorMedMCQA-V leaderboard

Source: Official KorMedMCQA-V leaderboard

#EntryScoreSettingDateLinkSource
1 Gemini 3 Pro 96.9% Reasoning Nov 2025 Official leaderboard
2 Gemini 3 Flash 96.9% Reasoning Dec 2025 Official leaderboard
3 GPT-5 93.9% Reasoning Aug 2025 Official leaderboard
4 GPT-5.2 93.9% Reasoning Dec 2025 Official leaderboard
5 GPT-5 mini 90.1% Reasoning Aug 2025 Official leaderboard
6 Qwen3-VL-32B-Thinking 83.7% Reasoning Oct 2025 huggingface.co Official leaderboard
7 Qwen3-VL-30B-A3B-Thinking 80.4% Reasoning Oct 2025 huggingface.co Official leaderboard
8 GLM-4.6V 78.7% Reasoning Dec 2025 huggingface.co Official leaderboard
9 Qwen3-VL-32B-Instruct 76.5% Oct 2025 huggingface.co Official leaderboard
10 Qwen3-VL-8B-Thinking 74.2% Reasoning Oct 2025 huggingface.co Official leaderboard
11 Ministral 3 14B 72.7% Reasoning Dec 2025 huggingface.co Official leaderboard
12 Qwen3-VL-30B-A3B-Instruct 71.7% Oct 2025 huggingface.co Official leaderboard
13 InternVL3.5-38B 67.2% Aug 2025 huggingface.co Official leaderboard
14 InternVL3.5-30B-A3B 66.4% Aug 2025 huggingface.co Official leaderboard
15 Lingshu-32B 32B 66.1% huggingface.co Official leaderboard
16 Qwen3-VL-4B-Thinking 65.9% Reasoning Oct 2025 huggingface.co Official leaderboard
17 GLM-4.6V-Flash 65.5% Reasoning Dec 2025 huggingface.co Official leaderboard
18 Ministral-3-8B-Reasoning-2512 8B 65.1% Reasoning huggingface.co Official leaderboard
19 Hulu-Med-32B 32B 64.7% huggingface.co Official leaderboard
20 Qwen2.5-VL-32B 63% Mar 2025 huggingface.co Official leaderboard
21 Gemma 3 27B 57.3% Mar 2025 huggingface.co Official leaderboard
22 MedGemma-27B-IT 27B 56.8% huggingface.co Official leaderboard
23 Ministral-3-14B-Instruct-2512 14B 56.5% huggingface.co Official leaderboard
24 Hulu-Med-14B 14B 55.1% huggingface.co Official leaderboard
25 Gemma 3 12B 54.3% Mar 2025 huggingface.co Official leaderboard
26 Qwen3-VL-8B-Instruct 53.8% Oct 2025 huggingface.co Official leaderboard
27 Ministral 3 8B Instruct 53.8% Dec 2025 huggingface.co Official leaderboard
28 InternVL3.5-14B 53.6% Aug 2025 huggingface.co Official leaderboard
29 InternVL3.5-8B 47.4% Aug 2025 huggingface.co Official leaderboard
30 Lingshu-7B 7B 46.6% huggingface.co Official leaderboard
31 Qwen2.5-VL-7B 46.6% Jan 2025 huggingface.co Official leaderboard
32 Hulu-Med-4B 4B 46.1% huggingface.co Official leaderboard
33 Qwen3-VL-4B-Instruct 45.7% Oct 2025 huggingface.co Official leaderboard
34 Ministral 3 3B Reasoning 45.2% Reasoning Dec 2025 huggingface.co Official leaderboard
35 MedGemma-1.5-4B-IT 4B 44.4% huggingface.co Official leaderboard
36 Hulu-Med-7B 7B 43.9% huggingface.co Official leaderboard
37 InternVL3.5-4B 43.3% Aug 2025 huggingface.co Official leaderboard
38 VARCO-VISION-2.0-14B 14B 43.2% huggingface.co Official leaderboard
39 A.X-4.0-VL-Light 7B 41.8% huggingface.co Official leaderboard
40 Ministral 3 3B Instruct 41.8% Dec 2025 huggingface.co Official leaderboard
41 Qwen3-VL-2B-Thinking 40.4% Reasoning Oct 2025 huggingface.co Official leaderboard
42 Kimi-VL-A3B-Thinking 37.6% Reasoning Apr 2025 huggingface.co Official leaderboard
43 MedGemma-4B-IT 4B 34.4% huggingface.co Official leaderboard
44 Gemma 3 4B 34.3% Mar 2025 huggingface.co Official leaderboard
45 Qwen3-VL-2B-Instruct 31.3% Oct 2025 huggingface.co Official leaderboard
46 Kimi-VL-A3B-Instruct 30.8% Apr 2025 huggingface.co Official leaderboard
47 Kanana-1.5-V-3B-Instruct 3B 30.7% huggingface.co Official leaderboard
48 InternVL3.5-2B 2B 29.8% huggingface.co Official leaderboard
49 HyperCLOVAX-SEED-Vision-Instruct-3B 3B 26.3% huggingface.co Official leaderboard
50 VARCO-VISION-2.0-1.7B 1.7B 24.4% huggingface.co Official leaderboard
51 InternVL3.5-1B 1B 22.4% huggingface.co Official leaderboard
#ModelScoreSettingDateLinkSource
1 Gemini 3 Pro 96.9% Reasoning Nov 2025 Official leaderboard
2 Gemini 3 Flash 96.9% Reasoning Dec 2025 Official leaderboard
3 GPT-5 93.9% Reasoning Aug 2025 Official leaderboard
4 GPT-5.2 93.9% Reasoning Dec 2025 Official leaderboard
5 GPT-5 mini 90.1% Reasoning Aug 2025 Official leaderboard
6 Qwen3-VL-32B-Thinking 83.7% Reasoning Oct 2025 huggingface.co Official leaderboard
7 Qwen3-VL-30B-A3B-Thinking 80.4% Reasoning Oct 2025 huggingface.co Official leaderboard
8 GLM-4.6V 78.7% Reasoning Dec 2025 huggingface.co Official leaderboard
9 Qwen3-VL-32B-Instruct 76.5% Oct 2025 huggingface.co Official leaderboard
10 Qwen3-VL-8B-Thinking 74.2% Reasoning Oct 2025 huggingface.co Official leaderboard
11 Ministral 3 14B 72.7% Reasoning Dec 2025 huggingface.co Official leaderboard
12 Qwen3-VL-30B-A3B-Instruct 71.7% Oct 2025 huggingface.co Official leaderboard
13 InternVL3.5-38B 67.2% Aug 2025 huggingface.co Official leaderboard
14 InternVL3.5-30B-A3B 66.4% Aug 2025 huggingface.co Official leaderboard
15 Qwen3-VL-4B-Thinking 65.9% Reasoning Oct 2025 huggingface.co Official leaderboard
16 GLM-4.6V-Flash 65.5% Reasoning Dec 2025 huggingface.co Official leaderboard
17 Qwen2.5-VL-32B 63% Mar 2025 huggingface.co Official leaderboard
18 Gemma 3 27B 57.3% Mar 2025 huggingface.co Official leaderboard
19 Gemma 3 12B 54.3% Mar 2025 huggingface.co Official leaderboard
20 Qwen3-VL-8B-Instruct 53.8% Oct 2025 huggingface.co Official leaderboard
21 Ministral 3 8B Instruct 53.8% Dec 2025 huggingface.co Official leaderboard
22 InternVL3.5-14B 53.6% Aug 2025 huggingface.co Official leaderboard
23 InternVL3.5-8B 47.4% Aug 2025 huggingface.co Official leaderboard
24 Qwen2.5-VL-7B 46.6% Jan 2025 huggingface.co Official leaderboard
25 Qwen3-VL-4B-Instruct 45.7% Oct 2025 huggingface.co Official leaderboard
26 Ministral 3 3B Reasoning 45.2% Reasoning Dec 2025 huggingface.co Official leaderboard
27 InternVL3.5-4B 43.3% Aug 2025 huggingface.co Official leaderboard
28 Ministral 3 3B Instruct 41.8% Dec 2025 huggingface.co Official leaderboard
29 Qwen3-VL-2B-Thinking 40.4% Reasoning Oct 2025 huggingface.co Official leaderboard
30 Kimi-VL-A3B-Thinking 37.6% Reasoning Apr 2025 huggingface.co Official leaderboard
31 Gemma 3 4B 34.3% Mar 2025 huggingface.co Official leaderboard
32 Qwen3-VL-2B-Instruct 31.3% Oct 2025 huggingface.co Official leaderboard
33 Kimi-VL-A3B-Instruct 30.8% Apr 2025 huggingface.co Official leaderboard

Rankings by part

We read 12 parts from the benchmark’s own leaderboard in 2 groups: modality and image count. Each ranks models on one part of the test. Gemini 3 Pro leads 5 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.

xray

33 models we track have a score here.

  1. 1 Gemini 3 Pro 97%
  2. 2 Gemini 3 Flash 96.2%
  3. 3 GPT-5.2 ↑1 94.2%
  4. 4 GPT-5 ↓1 93.6%
  5. 5 GPT-5 mini 89.1%
  6. 6 Qwen3-VL-32B-Thinking 83.9%
  7. 7 Qwen3-VL-30B-A3B-Thinking 80.7%
  8. 8 GLM-4.6V 78.4%
  9. 9 Qwen3-VL-32B-Instruct 75.9%
  10. 10 Qwen3-VL-8B-Thinking 75.1%

Number one over time

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

Google DeepMind
Own leaderboard Passed 85% in Feb 2026, by Gemini 3 Pro
Gemini 3 Pro 96.9%

Source: Official KorMedMCQA-V leaderboard

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
On the benchmark’s own leaderboard Best so far
0% 20% 40% 60% 80% 100% 2025 2026 Near the ceiling (85%) Gemini 3 Pro GPT-5 Qwen2.5-VL-32B Gemma 3 27B Qwen2.5-VL-7B 0% 20% 40% 60% 80% 100% 2025 2026 Near the ceiling (85%) Gemini 3 Pro Qwen2.5-VL-32B

Source: Official KorMedMCQA-V 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 8 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 2 pass 3 notes
  • Pass Independent runs Scores from the official leaderboard (51 entries).
  • Pass Models scored 33 models we track, plus 18 other entries on its leaderboard.
  • Note Newest score The newest came 11 months ago, in December 2025, for Gemini 3 Flash. 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-5.2 (#51).
  • Note Headroom Saturated. The best is 96.9%, past 85% since August 2025.
Leaderboard import 3 pass
  • Pass Importer We read the board on kormedmcqa-v.github.io: 33 rows for models we track and 18 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