Alibaba has launched Qwen 3.8-Max, an artificial intelligence model with 2.4 trillion total parameters including 95 billion active parameters, capable of processing context windows of approximately 1 million tokens. This multimodal model can ingest entire codebases, lengthy legal documents, or long reports in a single pass. Standard pricing is set at approximately $2 per million input tokens and $6 per million output tokens, with a 10% discount during the promotional period. Alibaba highlights the model’s performance in coding and agentic workflows, and plans to release open weights to allow developers to fine-tune it for specific needs. With 2.4 trillion parameters, Qwen 3.8-Max positions itself among the largest publicly acknowledged AI models in the world, using a mixture-of-experts architecture.
Source: Read the original article
