Multimodal Reasoning on MM-Vista

81.4Accuracy

GPT-5

Updated 25d ago

Evaluation Results

Method	Links
GPT-5 2026.06		81.4
Gemini-2.5-Pro 2026.06		80.9
Qwen3-VL-8B 2026.06		77.2
TACO (Ours-7B) 2026.06		76.3
MathCoder-VL-8B 2026.06		73.6
Qwen2.5-VL-32B 2026.06		72.2
DeepEyes-v2-7B 2026.06		71.9
CodeV-RL-7B 2026.06		71.8
PyVision-RL-7B 2026.06		71.4
Pixel-Reasoner-7B 2026.06		71.2
InternVL3-8B 2026.06		70.4
DeepEyes-7B 2026.06		70.1
Thyme-7B 2026.06		70
Qwen2.5-VL-7B 2026.06		68.2
Mini-o3-7B 2026.06		67.2
Unsilencing Latent Reasoning 2026.05		65.7
GPT-4o 2026.06		63.4
DMLR 2026.05		62.9
ICoT 2026.05		61.4
Vanilla 2026.05		61.1
CCoT 2026.05		60.5
MCoT 2026.05		59.6
LLaVA-OneVision-7B 2026.06		58.6
DMLR 2026.05		58
Unsilencing Latent Reasoning 2026.05		57.3
ICoT 2026.05		55.6
CCoT 2026.05		53.4
MCoT 2026.05		52.5
Vanilla 2026.05		51.2