Qwen released Qwen3.8-Max, 2.4T open-weight model
Qwen3.8-Max (2.4T parameters) matches Kimi K3 and DeepSeek V4 Flash across benchmarks, ranks #2 in Vision Arena behind Claude Fable 5 Max, and is stronger at coding. Open weights arrive next week alongside Qwen3.8-27B.
How this story unfolded
2 weeks · 0 reports · 8 community posts · 8 of 13 shown
- Jul 19
- Aug 3
Qwen by email
Get an email when Qwen has news
No news that day, no email.
More stories today
- Ante is a coding agent in a single binary that runs offline
- Demo converts static images into editable DrawIO XML via SAM 3
- Lindy reads Slack history and answers with source links
- VicOne releases free cybersecurity extension for NVIDIA Isaac Sim
- Curated collection tracks LLM-based software vulnerability detection