Skip to content

Benchmark

MedXpertQA (MM)

Scores are percentages. Higher is better.

Models measured
23
Scores recorded
31
Sources
3
IQ weight
2.15

Score distribution

65%
Mean
68.7%
Median
14%
Std dev
47.6%
10th percentile
80.4%
90th percentile
23.5%
Min
23.5%81.5%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.15 for chat models, measured over 23 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1GPT-5.6 Solclinical-result-548· Clinical Benchmarks data docs81.5%
2Gemini 3.1 Pro Previewclinical-result-156· Clinical Benchmarks data docs81.3%
3Qwen3.8 Maxclinical-result-157· Clinical Benchmarks data docs· Reported by the model's developer. The source relaying it did not verify it. claim80.4%
4Qwen3.8 MaxQwen3.8-Max release vision/video table · published setting· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.80.4%
5Claude Fable 5clinical-result-547· Clinical Benchmarks data docs80%
6Muse Sparkclinical-result-158· Clinical Benchmarks data docs· Reported by the model's developer. The source relaying it did not verify it. claim78.4%
7GPT 5.4clinical-result-159· Clinical Benchmarks data docs77.1%
8Gemini 3 Pro Previewclinical-result-1823· Clinical Benchmarks data docs76%
9GPT 5.2clinical-result-550· Clinical Benchmarks data docs73.3%
10Claude Opus 4.8clinical-result-546· Clinical Benchmarks data docs71.7%
11Qwen3.7 Plus· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim71%
12Qwen3.7 Plusclinical-result-160· Clinical Benchmarks data docs· Reported by the model's developer. The source relaying it did not verify it. claim71%
13Qwen3.7 PlusQwen3.7-Plus release · published setting· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.71%
14Qwen3.5 397B A17Bclinical-result-552· Clinical Benchmarks data docs· Reported by the model's developer. The source relaying it did not verify it. claim70%
15Qwen3.5 397B A17BQwen3.5 release/model card · published setting· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.70%
16Qwen3.6 Plusclinical-result-549· Clinical Benchmarks data docs· Reported by the model's developer. The source relaying it did not verify it. claim68.7%
17Grok 4.20clinical-result-161· Clinical Benchmarks data docs65.8%
18Kimi K2.5clinical-result-551· Clinical Benchmarks data docs65.3%
19Claude Opus 4.6clinical-result-162· Clinical Benchmarks data docs64.8%
20Claude Opus 4.5clinical-result-1824· Clinical Benchmarks data docs63.6%
21Gemma 4 31B ITclinical-result-1825· Clinical Benchmarks data docs· Reported by the model's developer. The source relaying it did not verify it. claim61.3%
22Gemma 4 31B ITGemma 4 Technical Report · Table 6 · thinking · 1120 vision tokens· Reported by Google DeepMind· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.61.3%
23Gemma 4 31B ITGemma 4 Technical Report · Table 12 · thinking · 280 vision tokens· Reported by Google DeepMind· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.60.7%
24Gemma 4 26B A4Bclinical-result-1826· Clinical Benchmarks data docs· Reported by the model's developer. The source relaying it did not verify it. claim58.1%
25Gemma 4 26B A4BGemma 4 Technical Report · Table 6 · thinking · 1120 vision tokens· Reported by Google DeepMind· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.58.1%
26Gemma 4 26B A4BGemma 4 Technical Report · Table 12 · thinking · 280 vision tokens· Reported by Google DeepMind· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.55.7%
27Qwen3.5 Omni PlusQwen3.5-Omni release · published setting· Reported by Qwen, Qwen Team· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.54.7%
28Qwen3 VL 235Bclinical-result-1827· Clinical Benchmarks data docs· Reported by the model's developer. The source relaying it did not verify it. claim47.6%
29Qwen3.5 Omni FlashQwen3.5-Omni release · published setting· Reported by Qwen, Qwen Team· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.44.8%
30gemma 4 E4Bclinical-result-1828· Clinical Benchmarks data docs· Reported by the model's developer. The source relaying it did not verify it. claim28.7%
31Gemma 4 E2B ITclinical-result-1829· Clinical Benchmarks data docs· Reported by the model's developer. The source relaying it did not verify it. claim23.5%

Data credits: Clinical Benchmarks (clinicalbenchmarks.ai), CC BY 4.0; fractions converted to percentages; source documents retain their own terms (licence) · BenchBoard by SYuan03 (github.com/SYuan03/benchboard). MIT covers its code and data organization; no results licence is inferred. Numeric relays retain the original stating-source links. · Data by LLM Stats