Qwen 3.8-Max-Prime

Changes from previous version

A faster serving lane for Qwen 3.8-Max, not a new checkpoint after Qwen 3.8-Flash-Next. Capabilities match Max; the difference is throughput at a higher token price.

Release Summary

High-throughput serving ID for Qwen 3.8-Max. Same 2.4T-parameter (95B active) weights, 1M-token context, and text, image, and video input. Alibaba Cloud international lists qwen3.8-max-prime at $4/M input, $0.50/M cache reads, and $12/M output, twice the $2/$6 rate recorded here for Qwen 3.8-Max. Published Max scores carry over, including 67.7 on SWE-bench Pro and 86.6 on Terminal-Bench 2.1.

Timeline

Other Qwen versions