Atlas

Benchmarks

← All benchmarks

Arena Text — Life, Physical & Social Science (No Style Control)

Professional Work

Raw Arena rating for life, physical, and social-science industry prompts without style-control adjustment. Higher ratings indicate stronger preference in blind pairwise user votes.

Top models (higher is better)

ModelScore
Gemini 4 Argon1542
Claude Fable 5.11533
Claude Opus 51531
Claude Opus 5.51525
Gemini 3.8 Flash1523
Opus 4.61519
Gemini 3.7 Flash1512
Qwen3.8-Max1512
Kimi K31508
Opus 4.71508
Muse Wonder1507
Muse Spark 1.31506
Claude Fable 51506
MiMo-V2.6-Pro1505
Muse Spark 1.21498
Gemini 3.1 Pro Preview1497
GLM-5.31497
Gemini 3.5 Flash1495
Gemini 3.6 Flash1489
GLM-5.21489
Claude Sonnet 5.51486
GPT-5.51485
Gemini 3 Pro Preview1485
ERNIE 5.11484
GLM-5.3 Flash1484
Muse Spark 1.11483
MiMo-V2.5-Pro1482
Gemini 2.5 Pro1480
Qwen3.7 Max Preview1479
GPT-5.41479
GLM-5.11478
Opus 4.81477
Muse Spark1475
Sonnet 4.61475
Gemini 3 Flash Preview1473
DeepSeek-V4-Pro1470
DeepSeek V4.1 Flash1470
Qwen3.8 27B1468
Kimi K2.61468
Nemotron 3 Ultra 550B A55B (NVFP4)1465
Loading Atlas data…