Atlas

Benchmarks

← All benchmarks

Arena Vision — Captioning (Style Controlled)

Multimodal

Style-controlled Vision Arena rating for image-captioning prompts. Higher ratings indicate stronger preference in blind pairwise user votes after adjusting for response style.

Top models (higher is better)

ModelScore
Gemini 3.1 Pro Preview1282
Gemini 3 Pro Preview1277
GPT-5.21262
Gemini 2.5 Pro1251
GPT-5.11248
Gemma 4 26B A4B IT1239
Gemini 3 Flash Preview1234
Gemini 3.1 Flash-Lite Preview1225
Gemini 2.5 Flash1211
GPT-4.11210
Kimi K2.51198
GPT-51196
GPT-4.1 Mini1193
O4 Mini1185
o31183
Gemini 2.5 Flash-Lite1182
GLM 5V Turbo1180
GPT-5 Mini1175
Qwen3.5 397B A17B1174
Grok 41170
Gemma 4 31B IT1170
Gemini 2.0 Flash1150
Mistral Small 3.1 24B Instruct 25031133
Mistral Medium 3.11130
Gemma 3 27B IT1117
Mistral Medium 31090
Mistral Small 3.2 24B Instruct 25061050
Loading Atlas data…