Alibaba launches Qwen 3.8 Max: 2.4T parameters, open weights promised

Alibaba officially released its flagship Qwen 3.8 Max model on a sparse MoE architecture with 2.4 trillion total parameters and 95 billion active per token. It is the first Max-tier Qwen promised with open weights, priced at $2 per million input and $6 per million output tokens.

• 2.4T total parameters, 95B active per token
• API pricing: $2/M input and $6/M output tokens
• Alibaba benchmarks show it beating GPT-5.6 Sol on 5 of 7 agent tasks
• Open weights promised "next week" after the August 3 release