Atlas

Benchmarks

← All benchmarks

Arena Text — Creative Writing (No Style Control)

Chat & Writing

Raw Arena rating for creative-writing 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 5.51534
Gemini 4 Argon1523
Claude Fable 5.11510
Opus 4.61505
Claude Opus 51495
Claude Fable 51493
Muse Wonder1493
Gemini 3.8 Flash1489
Gemini 3.7 Flash1489
Opus 4.71486
Gemini 3 Pro Preview1482
Gemini 3.1 Pro Preview1482
Qwen3.8-Max1477
Gemini 3.5 Flash1470
Claude Sonnet 5.51468
MiMo-V2.6-Pro1466
Gemini 3.6 Flash1466
Muse Spark 1.31462
GLM-5.21462
GLM-5.31459
Muse Spark1459
Gemini 3 Flash Preview1457
Kimi K31456
GPT-5.51455
Opus 4.81454
Gemini 2.5 Pro1454
GLM-5.11453
GPT-5.6 Sol1450
Qwen3.7 Max Preview1448
DeepSeek-V4-Pro1446
Opus 4.51443
Grok 4.51442
Sonnet 4.51442
ERNIE 5.11441
GLM-5.3 Flash1441
GPT-5.41439
DeepSeek V4 Pro 08131439
DeepSeek V4.1 Flash1438
Qwen3.7-Plus (source unspecified)1438
MiMo-V2.5-Pro1438
Loading Atlas data…