Qwen 3.8
Open weights
Alibaba
·
Alibaba’s Qwen3.8 generation (August 2026): the 27B checkpoint is Apache 2.0 with vision and a 262K context, and abliterated “uncensored” builds appeared within a day. The likely next base for roleplay fine-tunes.
Qwen3.8-27B shipped on 14 August 2026 under Apache 2.0 with 27.8B parameters, image and video input and a 262,144-token context; the 2.4T Max checkpoint shipped separately under a custom licence. Abliterated (refusal-removed) builds in FP8 and GGUF followed within a day, and roleplay fine-tunes are expected to follow the pattern of the 3.5 generation. MyBabes currently runs Qwen 3.5 writer tunes and is evaluating 3.8.
Skip the filter
Try it with adult content enabled
Where the vendor filters this model, MyBabes runs it uncensored inside its own generation stack. Generate from your own character images, no setup required.
Open MyBabes
From the Lab
Articles covering this model
All articles
News
Qwen3.8-27B Uncensored: Apache 2.0 Weights, the Abliterated Builds That Followed in a Day, and Whether It Replaces the 3.5 Roleplay Tunes
Qwen3.8-27B landed with an Apache 2.0 licence, vision input and a 262K context, and the "uncensored" versions were on Hugging Face the next day. Most of them are abliterations, not fine-tunes. Here is what that distinction means for adult roleplay, and when a real successor to the 3.5 writer tunes is likely.
·
3 min read
News
Meta Muse Glimmer: A 30B Apache 2.0 Model Built for Agents, and What It Would Take to Become a Roleplay Base
Meta is back to open weights with Muse Glimmer, a 30B model under Apache 2.0 that runs on one consumer GPU. It was built for tool use, not conversation, and it ships safety-tuned. Whether it ever matters for adult roleplay depends entirely on the fine-tuning community, so here is what they have to work with.
·
3 min read