Atlas

Benchmarks

← All benchmarks

Arena Text — Expert (No Style Control)

Chat & Writing

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

Top models (higher is better)

ModelScore
Claude Opus 51553
Claude Opus 5.51550
Gemini 4 Argon1545
Claude Fable 5.11545
Opus 4.61543
MiMo-V2.6-Pro1541
Claude Fable 51534
Claude Sonnet 5.51530
Gemini 3.8 Flash1527
Opus 4.71521
Kimi K31520
GPT-5.6 Sol1516
Muse Spark 1.31515
Muse Wonder1514
Gemini 3.7 Flash1514
GLM-5.31514
GLM-5.3 Flash1512
GPT-5.51507
GPT-5.41507
Qwen3.8-Max1507
Opus 4.81503
DeepSeek V4.1 Flash1502
MiMo-V2.5-Pro1500
Sonnet 4.61500
GPT-5.6 Terra1496
MiMo-V2.6-Flash1494
Gemini 3.5 Flash1494
GPT-6 Astra1492
ERNIE 5.11491
Sonnet 51490
Kimi K2.61490
Gemini 3.6 Flash1489
GPT-6.1 Sol1489
Qwen3.7 Max Preview1488
GLM-5.21487
Opus 4.51487
Gemini 3.1 Pro Preview1483
Qwen3.8 27B1480
GLM-5.11478
GPT-5.6 Luna1478
Loading Atlas data…