multimodal benchmark
A comprehensive benchmark to evaluate expert-level medical knowledge and advanced reasoning, featuring 4,460 questions spanning 17 specialties and 11 body systems. Includes both text-only and multimodal subsets with expert-level exam questions incorporating diverse medical images and rich clinical information.
Updated Aug 7, 2026
Sorted by the source-provided rank. Higher score is better according to the registry.
| 1 | ME | 78.4% | 100.0% | 12 | C | |
| 2 | AC | 67.3% | 90.9% | 12 | C | |
| 3 | AC | 62.4% | 81.8% | 12 | C | |
| 4 | AC | 61.4% | 72.7% | 12 | C | |
| 5 | GO | 61.3% | 63.6% | 12 | C | |
| 6 | GO | 58.1% | 54.5% | 12 | C | |
| 7 | GO | 49.0% | 45.5% | 12 | C | |
| 8 | GO | 48.7% | 36.4% | 12 | C | |
| 9 | MI | 43.0% | 27.3% | 12 | C | |
| 10 | GO | 28.7% | 18.2% | 12 | C | |
| 11 | GO | 23.5% | 9.1% | 12 | C | |
| 12 | GO | 18.8% | 0.0% | 12 | C |
Top published rows on the benchmark's original scale.
The top published results on this benchmark's own scale.
Definition and scoring fields from the benchmark registry.
A comprehensive benchmark to evaluate expert-level medical knowledge and advanced reasoning, featuring 4,460 questions spanning 17 specialties and 11 body systems. Includes both text-only and multimodal subsets with expert-level exam questions incorporating diverse medical images and rich clinical information.
Scores are shown in ratio. The current registry marks this benchmark as not independently verified with evidence level B.
Source-native results are preserved. Eligibility for the overall LLMBoard score is a separate policy decision.
Common questions about MedXpertQA.
Muse Spark is currently ranked first with 78.4%.
A comprehensive benchmark to evaluate expert-level medical knowledge and advanced reasoning, featuring 4,460 questions spanning 17 specialties and 11 body systems. Includes both text-only and multimodal subsets with expert-level exam questions incorporating diverse medical images and rich clinical information.
Yes. Higher values rank better for this benchmark.
12 unique published model results are currently shown.
This benchmark is preserved as source-native evidence but is not eligible for the current overall score.