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

Multimodal Reasoning on MathVista

89.8Pass@1

Gemini 3-Pro

Updated 25d ago

Evaluation Results

Method	Links
Gemini 3-Pro 2026.03		89.8
Seed1.8 2026.03		87.7
Seed 1.5-VL 2025.05		85.6
Seed1.5-VL Thinking 2026.03		85.6
Seed 1.5-VL 2025.05		83
Gemini 1.5 Pro 2025.05		82.7
Gemini 2.5-Pro 2026.03		82.7
GPT-5.1 High 2026.03		80.6
Claude Sonnet-4.5 2026.03		80.4
Qwen 2.5-VL 72B 2025.05		74.8
Claude 3.7 Sonnet 2025.05		74.5
RAP 2025.06		73.2
OpenAI o1 2025.05		71.8
LIMR 2025.06		71.1
Qwen2.5-VL-7b-Full 2025.06		70.7
LIMO 2025.06		69.9
Qwen2.5-VL-7b 2025.06		68.7
s1 2025.06		68.5
RAP 2025.06		64.9
Qwen2.5-VL-3b-Full 2025.06		64.5
GPT-4o 2025.05		63.8
Socratic-Solver-Geo (Stage3) 2026.02		63.55
GeoReasoning 2026.02		63.4
LIMR 2025.06		63.1
R-CoT 2026.02		62.6
s1 2025.06		62.6
KD (Our Synthesis) 2026.02		62.4
Socratic-Solver-Geo (Stage2) 2026.02		62.3
KD (Geo3K) 2026.02		62.1
Geo170k (G-LLaVA) 2026.02		62
TrustGeoGen 2026.02		61.8
LIMO 2025.06		61.8
PGPS9K 2026.02		61.3
Qwen2.5-VL-3b 2025.06		61.3
Socratic-Solver-Geo (Stage1) 2026.02		61.2
Qwen2.5-VL-7B-Instruct 2026.02		61.1