Atlas

Benchmarks

← All benchmarks

Arena Text — Writing, Literature & Language (No Style Control)

Professional Work

Raw Arena rating for writing, literature, and language industry prompts without style-control adjustment. Higher ratings indicate stronger preference in blind pairwise user votes.

Top models (higher is better)

ModelScore
Claude Opus 51511
Claude Fable 51508
Opus 4.61501
Opus 4.71490
Gemini 3.1 Pro Preview1479
Gemini 3.6 Flash1479
Gemini 3 Pro Preview1478
Kimi K31471
Gemini 3.5 Flash1469
GPT-5.51466
Opus 4.81465
Qwen3.7 Max Preview1459
GPT-5.6 Sol1459
GPT-5.41459
Gemini 3 Flash Preview1459
Muse Spark 1.11458
GLM-5.21456
GLM-5.11456
Gemini 2.5 Pro1454
Muse Spark1451
MiMo-V2.5-Pro1449
Opus 4.51449
DeepSeek-V4-Pro1448
ERNIE 5.11447
Sonnet 4.61445
Grok 4.51443
Qwen3.7-Plus1443
Sonnet 4.51441
GLM-51438
Kimi K2.61436
GPT-5.6 Terra1432
Sonnet 51429
Gemma 4 31B1429
Gemini 3.5 Flash-Lite1428
Kimi K2.51426
GPT-5.11426
MiMo-V2-Pro1424
GPT-5.6 Luna1422
MiniMax M31420
Qwen3.6 Plus (2026-04-02)1419
Loading Atlas data…