Atlas

Benchmarks

← All benchmarks

Hemingway-bench

Chat & Writing · 2026-02-04

Elo rating on real-world writing tasks graded by expert writers.

Top models (higher is better)

ModelScore
Claude Fable 51110
Gemini 3.7 Flash1105
Claude Fable 5.11095
Gemini 3.8 Flash1093
Gemini 3.6 Flash1075
Gemini 3.1 Pro Preview1073
Gemini 3.5 Flash1071
Kimi K31069
GPT-5.6 Sol1057
Gemini 3 Pro Preview1056
GLM-5.31053
Gemini 3 Flash Preview1052
Claude Opus 51047
Opus 4.61045
Opus 4.81044
Opus 4.71042
GLM-5.21036
GPT-5.51035
GLM-5.3 Flash1034
Kimi K2.61021
Opus 4.51016
Muse Spark 1.21013
Sonnet 4.61001
DeepSeek-V4-Pro999
GPT-5.2 Instant (source unspecified)994
Gemini 3.5 Flash-Lite990
Qwen3.5 Plus (2026-02-15)990
Kimi K2.5988
GPT-5.4985
Muse Spark 1.1985
Grok 4.6983
DeepSeek V4 Flash Vision Exp981
Grok 4.5972
Hy3971
DeepSeek-V4-Flash969
GPT-5.2954
Qwen3 Max (rolling alias)948
Grok 4.1 Fast916
Muse Glimmer 30B913
Nemotron 3.5 Lightning 30B A3B (precision unspecified)888
Loading Atlas data…