model-releasechinamultimodal Qwen3.8-Max is a 2.4T-param preview — here's what you can't test yet Qwen3.8-Max is a preview with no weights, no pricing, and no API — the 2.4T parameter count won't tell you what your pipeline actually needs.