Atlas

Benchmarks

← All benchmarks

Arena Vision — Captioning (No Style Control)

Multimodal

Raw Vision Arena rating for image-captioning prompts without style-control adjustment. Higher ratings indicate stronger preference in blind pairwise user votes.

Top models (higher is better)

ModelScore
Gemini 3.1 Pro Preview1345
Gemini 3 Pro Preview1325
Gemini 2.5 Pro1304
Gemma 4 26B A4B IT1297
Gemini 3 Flash Preview1281
Gemini 2.5 Flash1277
GPT-5.21267
GLM 5V Turbo1264
Gemini 3.1 Flash-Lite Preview1259
Kimi K2.51257
GPT-5.11252
Grok 41245
Qwen3.5 397B A17B1230
Gemma 4 31B IT1218
GPT-5 Mini1217
o31210
GPT-4.11207
O4 Mini1204
GPT-51204
Gemini 2.5 Flash-Lite1184
GPT-4.1 Mini1174
Mistral Small 3.1 24B Instruct 25031153
Gemma 3 27B IT1151
Mistral Medium 3.11145
Gemini 2.0 Flash1130
Mistral Medium 31094
Mistral Small 3.2 24B Instruct 25061059
Loading Atlas data…