List your company

Open-Style Question Answering on Image Implication Open-Style 1.0 (test)

3.94ScoreMetaphorStar-32B
1.46482.10742.753.3926Feb 11, 2026

Evaluation results

MethodLinks
MetaphorStar-32BParameters=32B2026.023.94
Gemini-3.0-pro2026.023.82
Claude-4.0-Sonnet2026.023.46
InternVL3-78BParameters=78B2026.023.42
Gemini-2.5-flash-thinking2026.023.42
Gemini-2.5-flash2026.023.34
GPT-4.12026.023.3
o4-mini2026.023.26
Grok-32026.023.24
Claude-3.5-Sonnet2026.023.22
MetaphorStar-7BParameters=7B2026.023.22
Doubao-1.5-thinking-vision-pro2026.023.16
QVQ-72BParameters=72B2026.023.1
QwenVL-2.5-32BParameters=32B2026.023.08
Grok-3-reasoning2026.023.06
MetaphorStar-3BParameters=3B2026.023.06
GLM-4V-plus2026.023.01
GPT-4o-mini2026.022.98
GPT-4o2026.022.94
DeepSeek-VL22026.022.82
GLM-4.1V-8BParameters=8B2026.022.6
QwenVL-2.5-3BParameters=3B2026.022.44
QwenVL-2.5-7BParameters=7B2026.022.34
LLaVA-1.5-7BParameters=7B2026.022.06
QwenVL-2.5-72BParameters=72B2026.021.56