Atlas

Benchmarks

← All benchmarks

LMArena Text Coding No Style Control Elo

Code · 2024-09-09

A Chatbot Arena text coding leaderboard that removes style-control adjustments and reports model Elo.

Top models (higher is better)

ModelScore
Claude Opus 5.51545
Gemini 4 Argon1539
Opus 4.61534
Claude Opus 51533
MiMo-V2.6-Pro1533
Claude Fable 5.11531
Opus 4.71518
Claude Fable 51518
Muse Wonder1518
Muse Spark 1.31511
Claude Sonnet 5.51510
GLM-5.3 Flash1508
Kimi K31508
Gemini 3.8 Flash1508
Muse Spark 1.21505
Sonnet 4.61503
Qwen3.8-Max1503
DeepSeek V4.1 Flash1501
MiMo-V2.5-Pro1501
Gemini 3.7 Flash1501
MiMo-V2.6-Flash1499
Opus 4.51498
GLM-5.31498
Muse Spark 1.11498
Qwen3.7 Max Preview1497
GPT-5.41497
GPT-5.6 Sol1496
GPT-5.51494
Gemini 3.5 Flash1493
Gemini 3.6 Flash1491
Opus 4.81491
GPT-6 Astra1489
Kimi K2.61488
ERNIE 5.11487
GLM-5.21487
GPT-5.6 Terra1486
GLM-5.11486
Sonnet 4.51485
Gemini 3.1 Pro Preview1484
Step 5 Preview1484
Loading Atlas data…