Atlas

Benchmarks

← All benchmarks

Arena Vision — English (Style Controlled)

Multimodal

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

Top models (higher is better)

ModelScore
Claude Fable 51303
Opus 4.71302
Opus 4.61301
Qwen3.8-Max1294
Gemini 3.8 Flash1292
Muse Spark1290
GPT-6 Astra1288
Gemini 3.7 Flash1287
GPT-5.6 Sol1284
Sonnet 4.61283
Opus 4.81281
Claude Opus 51281
GPT-5.41281
GPT-5.51281
GPT-6.1 Sol1279
Muse Spark 1.31279
Claude Fable 5.11279
Gemini 3.5 Flash1278
Muse Spark 1.11276
Gemini 3.6 Flash1276
Muse Spark 1.21275
Gemini 3 Pro Preview1275
Grok 4.51275
Gemini 3.1 Pro Preview1272
Step 5 Preview1271
GLM-5.3 Flash1270
GPT-5.6 Terra1269
DeepSeek V4.1 Flash1268
Grok 4.61267
Kimi K2.61267
Gemini 3.5 Flash-Lite1267
Claude Sonnet 5.51265
Qwen3.7-Plus (source unspecified)1264
Sonnet 51263
GPT-5.6 Luna1262
Gemini 3 Flash Preview1262
Gemma 4 31B IT1256
Kimi K2.51255
MiMo-V2.6-Flash1254
GPT-5.11251
Loading Atlas data…