SciEval-Leaderboard / Large Language Model Scientific Capability.csv
naonaowyh's picture
initial leaderboard
fc80ff8 verified
Raw
History Blame Contribute Delete
1.16 kB
Model,Type,Parameters,Knowl. Und.,Code Gen.,Symbolic Reason.,Hypoth. Gen.,Overall
Claude 4.5 Sonnet,Close,,60.67 ,21.73 ,40.36 ,56.10 ,44.72
Claude4-1-Opus,Close,,60.87 ,25.32 ,38.69 ,29.47 ,38.58
GPT-4o,Close,,60.84 ,17.67 ,32.09 ,33.04 ,35.91
GPT-5,Close,,74.05 ,29.21 ,39.91 ,45.67 ,47.21
GPT-o3,Close,,76.05 ,25.26 ,38.14 ,34.14 ,43.40
Gemini-2.5-Flash,Close,,50.46 ,18.28 ,32.07 ,40.86 ,35.42
Gemini-2.5-Pro,Close,,59.34 ,24.77 ,34.96 ,50.73 ,42.45
Grok-2-vision-1212,Close,,50.14 ,20.60 ,28.21 ,49.63 ,37.14
Ling-flash-2.0,Open,100B,53.39 ,25.60 ,37.98 ,50.29 ,41.81
Seed1.6-vision,Close,,65.78 ,21.49 ,39.24 ,45.00 ,42.88
DeepSeek-R1,Open,685B,45.17 ,0.06 ,20.00 ,49.73 ,28.74
GLM-4.5V,Open,106B,52.78 ,3.24 ,13.43 ,42.23 ,27.92
InternS1,Open,241B,66.14 ,17.08 ,31.62 ,37.45 ,38.07
Kimi-k2,Open,1040B,62.49 ,20.86 ,38.59 ,42.28 ,41.06
Llama 4 Maverick,Open,400B,57.22 ,18.26 ,38.97 ,38.31 ,38.19
Qwen3-VL-235B-A22B,Open,235B,65.98 ,18.00 ,49.93 ,40.62 ,43.63
Qwen3-Max,Open,1000B,63.14 ,43.97 ,41.04 ,42.12 ,47.57
GPT-5.1,Close,,69.23 ,25.63 ,32.44 ,41.45 ,42.19
Gemini-3-Pro,Close,,66.06 ,29.57 ,45.19 ,61.51 ,50.58