Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Multi-task Language Understanding on MMLU-Pro (Δ%)
Loading...
5.31
MMLU-Pro Delta (%)
MedGemma
-0.1916
1.2367
2.665
4.0933
Apr 20, 2026
MMLU-Pro Delta (%)
Updated 1mo ago
Evaluation Results
Method
Method
Links
MMLU-Pro Delta (%)
MedGemma
Size=4B, Variant=Med
2026.04
5.31
MedGemma
Size=27B, Variant=Med
2026.04
4.91
Qwen-2.5VL
Size=32B, Variant=FT
2026.04
2.25
Gemma-3
Size=4B, Variant=FT
2026.04
1.87
Gemma-3
Size=27B, Variant=FT
2026.04
1.04
Gemma-3
Size=12B, Variant=FT
2026.04
0.54
Qwen-2.5VL
Size=7B, Variant=FT
2026.04
0.02
Feedback
Search any
task
Search any
task