Atlas

Benchmarks

← All benchmarks

Arena Vision — Creative Writing (Style Controlled)

Multimodal

Style-controlled Vision Arena rating for visually grounded creative-writing prompts. Higher ratings indicate stronger preference in blind pairwise user votes after adjusting for response style.

Top models (higher is better)

ModelScore
Gemini 3 Pro Preview1325
Claude Fable 51319
GPT-6.1 Sol1311
Opus 4.61309
Muse Spark1303
Opus 4.71302
Gemini 3.1 Pro Preview1301
Gemini 3 Flash Preview1296
GPT-6 Astra1294
Gemini 3.7 Flash1292
Claude Fable 5.11290
Opus 4.81289
GPT-5.6 Sol1289
Qwen3.8-Max1288
GPT-5.51281
Grok 4.51281
Gemini 3.5 Flash1277
Gemini 3.8 Flash1276
Claude Opus 51275
Muse Spark 1.31272
Sonnet 4.61266
Gemma 4 31B IT1265
GPT-5.41262
Gemini 3.6 Flash1261
GPT-5.11259
Gemini 3.5 Flash-Lite1257
Grok 4.61257
GPT-5.6 Luna1257
Kimi K2.61254
DeepSeek V4.1 Flash1254
GPT-5.6 Terra1254
Gemini 2.5 Pro1251
Grok 4.31248
Kimi K2.51247
Gemini 2.5 Flash Preview (09-2025)1246
Qwen3.5 397B A17B1244
GLM-5.3 Flash1242
Qwen3.7-Plus (source unspecified)1239
Muse Spark 1.11238
Gemini 3.1 Flash-Lite Preview1237
Loading Atlas data…