Share your thoughts, 1 month free Claude Pro on usSee more

Long-form Writing Evaluation on WritingBench In-domain (test)

1,574Elo Rating

GPT-5.4

Updated 1mo ago

Evaluation Results

Method	Links
GPT-5.4 2026.06		1,574
Claude Opus 4.6 2026.06		1,468
Human 2026.06		1,464
POLARIS-9B 2026.06		1,419
Gemini 3.1 Pro 2026.06		1,387
Gemini 3 Flash 2026.06		1,335
Qwen3.5-27B 2026.06		1,285
Gemma-4-31B 2026.06		1,283
Gemma-4-26B-A4B* 2026.06		1,235
Gemma-4-E4B† 2026.06		1,108
Qwen3.5-9B 2026.06		1,086
Ministral-3-8B 2026.06		1,055
DeepSeek-R1-Distill-Qwen-14B 2026.06		891
LongWriter-Llama3.1-8B 2026.06		868
LongWriter-Zero-32B 2026.06		800