Hemmingway-1
A 27B-parameter open-weight language model, built on Qwen3.8-27B, tuned specifically for everyday writing such as messages and emails, released free for non-commercial use under a CC BY-NC license.
Hemmingway-1 is a 27-billion-parameter model built on Qwen3.8-27B and tuned to write everyday text directly, such as a message to a landlord or an overdue email, rather than returning several options with a preamble explaining them. The authors evaluated it against larger models on eighty everyday-writing requests, judged head to head without knowing which model produced which answer, and report that it beat every other model tested on that task, including much larger ones, and was also judged more often to read like something a person wrote.
The model is also reported to perform competitively on a public emotional-intelligence benchmark and to hold up reasonably in storytelling, though the authors note it loses to dedicated story-writing models on hostile storytelling and long narrative turns. It supports a context window of 262,144 tokens, serves through vLLM, and is released with open weights under a CC BY-NC 4.0 license, free for non-commercial use, with commercial use available by separate agreement; native Mac and Android apps are also offered alongside the raw weights.