← sources

we made a 27b model for creative writing. performs as good as claude fable 5, at a 40x cheaper price, open weights.

source: reddit (reddit:PromptEngineering)published 2026-09-20 13:23ingested 2026-09-21 02:38

retained text

hey! we're a small lab in switzerland and we just put out hemmingway-1. it's a 27b model we trained specifically for creative writing, dialogue, roleplay, essays, and for everyday texting. frontier performance at a 40x cheaper price.

we thought it might be useful for people here who write prompts and want a model that writes well.

it benchmarks at absolute frontier level. on eq-bench 4 it scores 1330, right behind claude fable 5, and beating many other frontier models including gpt-5.5 and opus 4.8. you can see our internal benchmarks on the model card: hemmingway.io/model/

the testers we gave it to told us it's the best ai they've ever used for writing and that it genuinely sounds human.

the weights are public, so you can run it yourself.

we also built an app for it. its autopilot function answers your emails and texts for you. mac, windows, android.

try it out for free at hemmingway.io!

document fetched 2026-09-21 00:35 · canonical https://reddit.com/r/PromptEngineering/comments/1wlh0mv/we_made_a_27b_model_for_creative_writing_performs