Atlas

Benchmarks

← All benchmarks

Arena Text — English (No Style Control)

Chat & Writing

Raw Arena rating for English-language Text Arena prompts without style-control adjustment. Higher ratings indicate stronger preference in blind pairwise user votes.

Top models (higher is better)

ModelScore
Opus 4.61513
Claude Opus 51502
Claude Fable 51498
Opus 4.71493
Gemini 3.6 Flash1484
Gemini 3 Pro Preview1480
Muse Spark1480
Gemini 3.5 Flash1479
Muse Spark 1.11479
GLM-5.11479
MiMo-V2.5-Pro1478
Gemini 3.1 Pro Preview1478
Kimi K31477
ERNIE 5.11477
Sonnet 4.61473
Qwen3.7 Max Preview1472
GLM-5.21471
GPT-5.41468
GPT-5.51468
Opus 4.81467
Gemini 3 Flash Preview1465
Qwen3.7-Plus1461
Opus 4.51460
Kimi K2.61458
Gemini 2.5 Pro1457
DeepSeek-V4-Pro1457
GPT-5.6 Sol1457
GLM-51457
Grok 4.51455
GLM-4.71454
Kimi K2.51454
MiMo-V2-Pro1454
GLM 5V Turbo1454
Sonnet 51453
Sonnet 4.51451
MiniMax M31451
Hy31451
GPT-5.11449
GLM 4.61448
Gemma 4 31B1448
Loading Atlas data…