Alibaba releases Qwen3.8-Max, a 2.4T-parameter open-weights model

Qwen3.8-Max, a sparse MoE with 2.4T total parameters and 256K context, scores 56-58 on the Artificial Analysis Intelligence Index at $1.76 per task, trailing Kimi K3 ($0.86) but rivaling Claude Sonnet 5. Open weights promised next week, with 27B variant.