Atlas

Benchmarks

← All benchmarks

Arena Text — German (No Style Control)

Chat & Writing

Raw Arena rating for German-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 3 Pro Preview1516
Opus 4.61508
Muse Spark1494
Opus 4.71494
Gemini 3.5 Flash1494
Gemini 3 Flash Preview1494
Gemini 3.1 Pro Preview1492
Claude Fable 51486
Gemini 2.5 Pro1486
Qwen3.7-Plus1480
GPT-5.6 Luna1478
GPT-5.41468
GPT-5.51468
GLM-5.11467
Qwen3 Max Preview1466
GLM-5.21466
ERNIE 5.11465
Muse Spark 1.11463
Opus 4.81460
DeepSeek-V4-Pro1457
Qwen3.6 Plus (2026-04-02)1455
Grok 4.51453
GPT-5.6 Sol1451
GLM 4.61450
Kimi K2.61450
MiMo-V2.5-Pro1449
Qwen3.5 397B A17B1449
GPT-5.21446
Grok 4.11445
GPT-5.6 Terra1443
Opus 4.51442
Kimi K2.51438
GPT-5.11438
Qwen3-Max (2025-09-23)1438
GLM-51437
Qwen3 235B A22B Instruct 25071436
Inkling1436
DeepSeek-V3.2-Exp1434
Sonnet 51434
Mistral Medium 3.11430
Loading Atlas data…