Atlas

Benchmarks

← All benchmarks

Arena Vision — Chinese (Style Controlled)

Multimodal

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

Top models (higher is better)

ModelScore
Gemini 3.8 Flash1384
Gemini 3.7 Flash1369
Muse Spark 1.31366
Claude Opus 51363
GPT-6 Astra1358
Claude Fable 51355
Opus 4.61352
Claude Fable 5.11351
Qwen3.8-Max1350
Opus 4.71347
Gemini 3.5 Flash1346
Gemini 3 Pro Preview1343
Gemini 3.1 Pro Preview1339
GLM-5.3 Flash1338
Muse Spark 1.11335
Gemini 3.6 Flash1335
Opus 4.81334
Muse Spark1334
Grok 4.51330
GPT-5.51329
Kimi K2.61326
GPT-5.41316
Sonnet 51316
Gemini 3 Flash Preview1314
GPT-5.6 Terra1313
Qwen3.7-Plus (source unspecified)1312
MiMo-V2.51312
DeepSeek V4.1 Flash1311
Gemma 4 31B IT1305
GPT-5.6 Sol1305
Kimi K2.51305
Gemini 3.5 Flash-Lite1303
Sonnet 4.61302
GPT-5.21300
Qwen3.8 27B1292
Qwen3.5 397B A17B1291
Gemma 4 26B A4B IT1290
MiniMax M31289
Grok 4.61287
GPT-5.4 Mini1285
Loading Atlas data…