List your company

Multimodal Reasoning and Perception on MMStar (test)

63.9AccuracyGPT-4o
51.31654.58357.8561.117Apr 6, 2026

Evaluation results

MethodLinks
GPT-4o2026.0463.9
InternVL2.5-8B2026.0463.2
Qwen2.5-VL-7B + Saliency-R1Training Stage=Saliency-R12026.0462.6
Qwen2.5-VL-7BTraining Stage=Base2026.0462.4
Claude-3.5 Sonnet2026.0462.2
LLaVA-OneVision-7B2026.0461.9
InternVL2-8B2026.0461.5
Insight-V-8B2026.0461.5
Vision-R1-LlamaV-CI-11B2026.0461.4
Mulberry-7B2026.0461.3
Qwen2.5-VL-7B + SFTTraining Stage=SFT2026.0460.1
IXC-2.52026.0459.9
Gemini-1.5-Pro2026.0459.1
MiniCPM-V-2.6-8B2026.0457.5
Qwen2.5-VL-3BTraining Stage=Base2026.0456.1
Qwen2.5-VL-3B + Saliency-R1Training Stage=Saliency-R12026.0456
Qwen2.5-VL-3B + SFTTraining Stage=SFT2026.0455.7
LLaVA-CoT-11B2026.0454.9
MiniCPM-Llama-V-2.5-8B2026.0451.8