Qwen3.8-27B open weights released, scores 52 on AI Index

Alibaba's Qwen3.8-27B, a 27B-parameter multimodal dense model, is now open-weight and outperforms Qwen3.7-Plus overall. It scores 52 on the Artificial Analysis Intelligence Index, matching GPT-5.6 Luna (max) and one point behind GLM-5.2 (max) and DeepSeek V4 Pro 0813 (max).
How this story unfolded
11 days · 8 reports · 36 community posts · from Aug 14
- Aug 14
- Aug 15
- Aug 16
Blackfrost-AI/Qwen3.8-27B-ABLITERATED-GGUFhuggingface.co
AEON-7/Qwen3.8-27B-AEON-ULTIMATE-UNCENSORED-BF16huggingface.co
Qwen 3.8 27B is excellent, but it defaults to wildly overthinking thingssimonwillison.net
Here’s my review of Qwen 3.8 27B - I can't remember the last time I've had this much fun playing...
- Aug 17
- Aug 18
- Aug 19
- Aug 20
- Aug 22
- Aug 23
- Aug 24
- Aug 25
Qwen by email
Get an email when Qwen has news
No news that day, no email.
More stories today
- Users question OpenAI's 30-day chat deletion promise
- 80% of developers find AI coding more addictive than helpful
- User asks ChatGPT to define a hyper-large number, 'Kevinber'
- Parag Agrawal: Building a New Web for AI Agents
- OpenAI product head discusses ChatGPT Work, Codex, and reporting to Greg Brockman