Qwen3.8-Max

Qwen3.8-Max is Alibaba's current Max-tier Qwen model for complex reasoning, coding, visual understanding, research, and long-running agents. The hosted model ID is qwen3.8-max. Current provider listings document a 1M-token context window, up to 131,072 output tokens, multimodal input, thinking and non-thinking modes, function calling, structured output, and built-in tools. OpenRouter lists standard provider pricing of $2 per million input tokens and $6 per million output tokens.

Official sources: Alibaba Model Studio model information, Alibaba model pricing, and the Qwen3.8-Max provider listing.

Qwen3.8-Max succeeds the Qwen3.8 Max Preview and is Alibaba's highest-capability hosted Qwen model at this research cutoff. Alibaba positions it for coding, research, office workflows, native visual understanding, long documents and videos, and autonomous multi-step execution.

The hosted model and the downloadable Qwen3.8 checkpoint must be evaluated separately. The hosted route includes native visual input, non-thinking mode, a default 1M context window, and official built-in tools. The open checkpoint gives teams weight-level control, but does not automatically inherit every hosted feature.

• Repository-scale coding and debugging
• Long-running research and agent workflows
• Multimodal document, screenshot, and video analysis
• Tool-connected business workflow automation
• Complex planning that must survive many iterative steps

The 2.4T model is expensive to self-host at full precision, provider-reported benchmark and autonomy claims are not independent success rates, and hosted/open behavior can differ. Validate the exact route, prompt, reasoning settings, and tools before production use.