Share your thoughts, 1 month free Claude Pro on usSee more

Geometric Reasoning on GeoAux-Bench Core (Total)

83.16Accuracy

Gemini-2.5-Pro

Updated 1mo ago

Evaluation Results

Method	Links
Gemini-2.5-Pro 2026.03		83.16
GPT-5 2026.03		80.62
Gemini-2.5-Flash 2026.03		79.53
Qwen3-VL-235B-Ins. 2026.03		74.85
Qwen3-VL-30B-Ins. 2026.03		70.25
Doubao-seed-1.6 2026.03		65
Qwen3-VL-235B-Thk. 2026.03		62.25
Qwen3-VL-30B-Thk. 2026.03		58.75
Claude-Opus-4.1 2026.03		55.82
Claude-Sonnet-4.5 2026.03		55.03
MiMo-VL-7B-SFT 2026.03		50.87
MiMo-VL-7B-RL 2026.03		50.7
GLM-4.5V 2026.03		40.24
InternVL3.5-8B 2026.03		36.26
InternVL3-8B 2026.03		35.17
Qwen2.5-VL-Ins. 2026.03		33.25
GLM-4.1V-Thk. 2026.03		31.76
GPT-4o 2026.03		28.13
MathCanvas-7B 2026.03		24.63
BAGEL-7B-MoT 2026.03		12.95
BAGEL-Zebra-CoT 2026.03		12.03