25 Sept 2026
Will social media survive the AI change? What the evidence says
Our overnight sample had no social referrers, while Meta's crawlers made 5.9% of page requests. Usage, labelling rules and the EU AI Act…
A 27B open-weight fine-tune of Qwen3.8-27B for emails, messages and short stories, with a 262,144-token context; weights under CC BY-NC 4.0 since 22 September, hosted API $0.24 in and $0.90 out per million tokens.
Not yet independently verified. The developer's identity is unclear: the weights sit under the Hugging Face organisation Altworld, the site brands itself Hemmingway, and its terms name the Commons Lab. Benchmark results are the developer's own, and its EQ-Bench 4 placing was not in the public leaderboard data on 22 September 2026. We will update this when it can be confirmed, and remove this note.
Hemmingway-1 is a 27-billion-parameter open-weight language model tuned for everyday writing: emails, text messages, replies, letters and short stories. It is a fine-tune of Alibaba's Qwen3.8-27B. The weights are published by the Altworld organisation on Hugging Face and the model is marketed at hemmingway.io, whose terms say the service is made and run by the Commons Lab, a small lab in Switzerland and South Africa. No vendor release date is given; the Hugging Face repository was created on 20 September 2026, and third-party quantised copies appeared the same day.
The model has a 262,144-token context and writes replies of up to 32,768 tokens including its reasoning, which is on by default. It reads images and supports OpenAI-style function calling. It is English-first. The weights were first published under Apache 2.0, but the licence on the Hugging Face repository was changed at 16:48 UTC on 22 September 2026 to CC BY-NC 4.0, with commercial use by agreement with the developer. A business that wants to use the model commercially now needs that agreement, and one that downloaded an earlier copy should check which licence that copy was published under before relying on it.
A hosted, OpenAI-compatible API (model id hemmingway-27b) costs $0.24 per million input tokens, $0.024 per million cached input tokens and $0.90 per million output tokens, with up to eight concurrent requests per account. Consumer apps for Mac, Windows, Android and the web start at $9 a month, with a free trial. The desktop and Android apps are not signed by Apple or Microsoft or listed in the Play Store, so installing them means overriding security warnings.
The developer's headline results come from three benchmarks it built and ran itself (CommunicationBench, Human-Likeness and StoryBench), which it discloses. It also reports third place on the public EQ-Bench 4, run with the official harness; Hemmingway-1 did not appear in the public EQ-Bench 4 leaderboard data on 22 September 2026. The developer warns that the model can be confidently wrong and should not be used for medical, legal or financial decisions.
| Benchmark | Official | Community avg |
|---|---|---|
| No benchmark scores yet. Be the first to add one. | ||
“Official” values are editor-approved and feed the ranking. “Community avg” is the mean of member submissions (shown for transparency; it never affects the ranking until an editor approves a value).
Sign in to add a benchmark score for this model.
An earned signal from verification, reviews, awards, transparency and engagement — the vendor can't buy it.
Updated 9/22/2026
Used it? Your experience helps other buyers decide.
Write a reviewNo questions yet. Be the first to ask about Hemmingway-1.
Everything here links back to the same verified catalogue. Pick your next stop.