Alibaba released Qwen3.8-Max, a 2.4T-parameter flagship AI model

Sparse 2.4T-parameter MoE priced at $2/$6 per million tokens, claiming to rival Claude Fable 5 and GPT-5.6 Sol Max on agentic computer use. Open weights ship next week, a first for Qwen's Max line, with day-zero availability on Together AI and Vercel AI Gateway. Artificial Analysis scores it 58, behind open-weights leader Kimi K3.