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

Multimodal Math Reasoning on WeMath

98.7Accuracy

Updated 1mo ago

Evaluation Results

Method
o1 2026.03	98.7	-
o1 2026.03	98.7	-
Qwen3.5-27B 2026.04	84	-
MM-Eureka-32B-R-TAP 2026.03	79.3	-
EXAONE 4.5 33B 2026.04	79.1	-
Qwen2.5-VL-7B-Instruct + R-TAP 2026.03	78.2	-
Gemini-2.5-Pro-Thinking 2025.12	78	-
Seed-1.5-thinking 2025.12	78	-
Gemini-2.5-Pro 2025.09	78	-
Qwen3-VL-235B-A22B 2026.04	74.8	-
CEPO 2026.05	74.77	-
Base 2026.05	74.31	-
SDPO 2026.05	73.62	-
MM-Eureka-32B 2026.03	73.4	-
RLSD 2026.05	73.28	-
GRPO 2026.05	73.1	-
ToR-DAPO 2026.03	73	-
Claude3.7-Sonnet 2026.03	72.6	-
Claude3.7-Sonnet 2026.03	72.6	-
Qwen-2.5-VL-72B 2026.03	72.4	-
Qwen-2.5-VL-72B 2026.03	72.4	-
OPSD 2026.05	72.36	-
ToR-DAPO 2026.03	72.1	-
Qwen2.5-VL-7B-Instruct + NoisyRollout 2026.03	71.9	-
Qwen3-VL-32B 2026.04	71.6	-
Gemini2-flash 2026.03	71.4	-
Gemini-2-flash 2026.03	71.4	-
GPT-5-Thinking 2025.09	71.1	-
MM-Eureka-7B-R-TAP 2026.03	71	-
GPT-5 mini 2026.04	70.3	-
NoisyRollout-7B 2026.03	69.6	-
DIVA-GRPO-7B 2026.03	69.3	-
DAPO 2026.03	69.3	-
ThinkLite-7B-VL 2026.03	69.2	-
ThinkLite-7B-VL 2026.03	69.2	-
Qwen-2.5-VL-32B 2026.03	69.1	-
Qwen-2.5-VL-32B 2026.03	69.1	-
ToR-GRPO 2026.03	68.9	-
GPT-4o 2026.03	68.8	-
GPT-4o 2026.03	68.8	-
Qwen2.5-VL-7B-Instruct + Vanilla GRPO 2026.03	68.6	-
R1-ShareVL-7B 2026.03	67.9	-
VLAA-Thinker-Qwen2.5-7B 2026.03	67.9	-
VLAA-Thinker-7B 2026.03	67.9	-
OpenVLThinker-7B 2026.03	67.8	-
OpenVLThinker-7B 2026.03	67.8	-
InternVL2.5-VL-38B 2026.03	67.5	-
InternVL2.5-VL-38B 2026.03	67.5	-
GRPO 2026.03	67.4	-
ADORA-7B 2026.03	67.1	-
InternVL2.5-VL-78B 2026.03	66.3	-
InternVL2.5-VL-78B 2026.03	66.3	-
InternVL2.5-38B-MPO 2026.03	66.2	-
InternVL2.5-38B-MPO 2026.03	66.2	-
MM-Eureka-7B 2026.03	66.1	-
MM-Eureka-7B 2026.03	66.1	-
MM-Eureka-Qwen-7B 2026.03	65.5	-
MM-Eureka-Qwen-7B 2026.03	65.5	-
QVQ-72B-Preview 2026.03	65.4	-
QVQ-72B-Preview 2026.03	65.4	-
Metis 2026.04	65.2	-
OpenVLThinker-7B 2026.03	64.3	-
OpenVLThinker-7B 2026.03	64.3	-
Adora-7B 2026.03	64.2	-
ADORA-7B 2026.03	64.2	-
Qwen2.5-VL-7B-Instruct 2026.03	63.1	-
Qwen2.5-VL-7B-Instruct 2026.03	63.1	-
Claude-4-Sonnet 2025.12	63	-
ToR-DAPO 2026.03	62.4	-
Qwen-2.5-VL-7B 2026.03	62.1	-
Qwen-2.5-VL-7B 2026.03	62.1	-
R1-OneVision-7B 2026.03	62.1	-
R1-OneVision-7B 2026.03	62.1	-
R1-Onevision-7B 2026.03	61.8	-
R1-Onevision-7B 2026.03	61.8	-
ToR-DAPO 2026.03	59.6	-
DAPO 2026.03	57.8	-
SFT-7B 2026.03	57.6	-
InternVL3.5-8B 2026.06	57	-
Gemini-2.0-Pro 2025.06	56.5	-
CEPO 2026.05	56.21	-
GPT-4.1 2025.12	55.5	-
GPT-4.1-20250414 2025.09	55.5	-
RLSD 2026.05	55.26	-
GRPO 2026.05	54.6	-
Ours 2025.12	54.4	-
InternVL2.5-VL-8B 2026.03	53.5	-
InternVL2.5-8B-MPO 2026.03	53.5	-
InternVL2.5-VL-8B 2026.03	53.5	-
InternVL2.5-8B-MPO 2026.03	53.5	-
Base 2026.05	52.24	-
Qwen2.5VL-72B x Qwen3-32B 2025.09	51.8	-
ChatGPT-4o-latest 2025.09	50.6	-
GPT-4o 2025.06	50.6	-
Claude3.7-Sonnet 2025.09	49.3	-
Claude-3.7-Sonnet 2025.06	49.3	-
Qwen2.5-VL-72B 2025.09	49.1	-
Qwen2.5-VL-72B 2025.06	49.1	-
InternVL3.5-4B (+GNDPO) 2026.06	48.8	-
MathBook 2025.12	48.4	-

Showing 100 of 238 rows