Qwen3.8-27B open-weight model launches, tops Hugging Face

Alibaba's Qwen3.8-27B, a 27B-parameter multimodal dense model with 262K native context, outperforms Qwen3.7-Plus overall. It scores 52 on the Artificial Analysis Intelligence Index, matching GPT-5.6 Luna (max), and is #1 trending on Hugging Face.
How this story unfolded
4 weeks · 3 reports · 47 community posts · 50 of 52 shown
- Aug 3
- Aug 14
- Aug 15
- Aug 16
- Aug 17
- Aug 18
- Aug 19
- Aug 20
- Aug 21
- Aug 23
- Aug 24
- Aug 25
- Aug 26
- Aug 27
- Aug 28
- Aug 29
- Aug 30
Qwen by email
Get an email when Qwen has news
No news that day, no email.
More stories today
- DeepMind panel discusses generative media SOTA and human eval
- Reddit users share useful MCP servers for Claude
- Why embodied AI hits an edge AI wall requiring new math
- HTMX CEO mandates 'No AI Fridays' to counter LLM cognitive debt
- OpenAI product lead Tara Seshan discusses persistent AI coworkers