Hemmingway Try Hemmingway

Blog · Guide

The best open-source LLM for writing

Hemmingway-1 is a 27B open-weights model under Apache-2.0, built for writing that sounds like a person. How it scores, what it needs, and how to run it.

By the Hemmingway team ·

If you want a writing model you can run yourself, fine-tune, or build a product on without asking anyone, you need open weights and a licence that lets you use them. Hemmingway-1 has both, and it was built for one job: writing that sounds like a person wrote it.

What it is

Parameters27B
Built onQwen3.8-27B
Context262,144 tokens
LicenceApache-2.0, including commercial use
Weightshuggingface.co/Altworld/Hemmingway-1
Codegithub.com/lukeckprobierts/Hemmingway-1

How it scores against open and closed models

The most telling comparison for an open model is against the model it started from. Hemmingway-1 is Qwen3.8-27B trained further for writing, so the base model's scores show what the training did.

TestHemmingway-1Qwen3.8 27B (base)
CommunicationBench (everyday writing)1026954
Human-Likeness1032952
StoryBench (creative writing)1197693

Against much larger models, on everyday writing it came first of all we tested, level with Fable 5.1 and ahead of Fable 5, GLM-5.3, Kimi K3, GPT-6 Astra, Grok 4.6 and DeepSeek V4 Pro. On the public EQ-Bench 4 it placed third. On long stories it sits level with Kimi K3, ahead of Qwen3.8 Max and DeepSeek V4 Pro, and behind Fable 5 Max and GLM-5.3. All of the scores are on the model card.

CommunicationBench, Human-Likeness and StoryBench are our own benchmarks. Every matchup was blind, run in both orders, and judged by a model that was not one of those being judged. EQ-Bench 4 is not ours.

Running it

With vLLM:

vllm serve Altworld/Hemmingway-1 --max-model-len 262144

Or with Transformers, as a normal chat model: load Altworld/Hemmingway-1 with AutoModelForCausalLM and AutoTokenizer, and build the prompt with the tokenizer's chat template. The README on GitHub has a full example.

Or skip the hardware

If you would rather not run a 27B model yourself, the same model is behind an OpenAI-compatible API as hemmingway-27b. Most code that already talks to a model needs two lines changed. The API docs are here, and keys are on the API platform.

What it is not good at

It is English-first. It can be wrong and still sound certain, so do not use it to decide anything medical, legal or financial. And for very long fiction, the largest story models are better.

Read next

  • The best AI for creative writing in 2026Which AI models write the best stories, poems and letters? Our blind StoryBench and EQ-Bench 4 results, including where our own model loses.
  • EQ-Bench 4: how Hemmingway-1 scoresEQ-Bench 4 tests emotional intelligence in conversation. We ran it on Hemmingway-1 with the official harness: 1330, third of 27, ahead of GPT-5.5 and Opus.
  • The best AI for writing emails and messagesWe tested the largest AI models on eighty real email and message requests in blind head-to-head matchups. Here are the scores, and what they mean.
  • Hemmingway vs Claude for writingClaude is one of the best writers among AI models. How Hemmingway-1 compares with Fable 5.1, Fable 5 and Opus on emails, sounding human, emotional intelligence and stories.