Atlas

Benchmarks

← All benchmarks

Arena Vision — Diagram (No Style Control)

Multimodal

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

Top models (higher is better)

ModelScore
Claude Fable 51370
Claude Opus 51341
Opus 4.71339
Opus 4.61339
GPT-5.51328
GPT-5.41323
Grok 4.51322
Sonnet 51319
Opus 4.81318
GPT-5.6 Terra1314
Gemini 3.1 Pro Preview1312
Muse Spark1310
Gemini 3.5 Flash1309
Sonnet 4.61305
Muse Spark 1.11301
Gemini 3 Pro Preview1301
Kimi K2.61297
GPT-5.6 Sol1297
Gemini 3 Flash Preview1291
GLM 5V Turbo1284
Gemma 4 31B1281
Qwen3.7-Plus1281
Kimi K2.51281
Qwen3.5 397B A17B1281
GPT-5.6 Luna1279
MiMo-V2.51276
Gemini 2.5 Pro1272
MiniMax M31269
GPT-5.4 Mini1264
Gemma 4 26B A4B1264
GPT-5.21264
GPT-5.11262
MiMo-V2-Omni1255
Qwen3.5 122B A10B1251
Gemini 2.5 Flash Preview (09-2025)1249
Gemini 2.5 Flash1247
Grok 4.31243
Gemini 3.1 Flash-Lite Preview1241
GPT-51228
O3 (2025-04-16)1218
Loading Atlas data…