Qwen 3.8-2.4T-A95B

Released on August 12, 2026

Changes from previous version

The self-hostable weights behind Qwen3.8-Max. Drops hosted-only vision, 1M default context, and built-in tools for a text MoE you can run locally.

Release Summary

Open-weight checkpoint of Qwen3.8-Max: 2.4T total / 95B active sparse MoE, text-only, 262K native context (extensible toward 1M) under the Qwen3.8-Max License. Published on Hugging Face and ModelScope as Qwen3.8-2.4T-A95B.

Timeline

August 12, 2026

Open weights published

Qwen publishes Qwen3.8-2.4T-A95B on Hugging Face and ModelScope as the first Max-class open-weight Qwen model.