Atlas

Benchmarks

← All benchmarks

Arena Text — Chinese (No Style Control)

Chat & Writing

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

Top models (higher is better)

ModelScore
Gemini 4 Argon1600
Claude Fable 5.11584
Claude Opus 5.51581
Claude Opus 51573
Gemini 3.7 Flash1553
Gemini 3.8 Flash1553
Opus 4.61551
Claude Fable 51544
Muse Wonder1541
Qwen3.8-Max1539
Claude Sonnet 5.51535
GPT-5.51534
GLM-5.31532
GLM-5.3 Flash1531
Opus 4.71531
Gemini 3.6 Flash1531
MiMo-V2.6-Pro1530
Kimi K31529
Gemini 3.1 Pro Preview1529
Muse Spark 1.31529
Qwen3.7 Max Preview1527
Gemini 3.5 Flash1527
GPT-5.6 Sol1526
Gemini 3 Pro Preview1523
GLM-5.21522
Kimi K2.61520
GPT-5.41519
Muse Spark 1.11515
GLM-5.11514
GPT-5.6 Terra1514
GLM-51511
Gemini 3 Flash Preview1510
Qwen3.7-Plus (source unspecified)1509
Muse Spark1508
MiMo-V2.6-Flash1507
Gemini 2.5 Pro1507
ERNIE 5.11507
Opus 4.81506
MiMo-V2.5-Pro1506
Muse Spark 1.21506
Loading Atlas data…