Qwen 3.7 Max is now available on ChatBotKit. Alibaba's newest flagship Qwen model is built specifically for the agent-centric era - emphasising long-term autonomous execution, strong programming output, and the kind of broad capability coverage that lets a single model carry a full agent workflow from start to finish.
Where Qwen 3.6 Max focused on refining coding performance and front-end development quality, Qwen 3.7 Max extends that foundation into sustained, multi-step execution. The model is designed to stay effective across extended task chains - retaining context, making better decisions mid-workflow, and handling the kind of productivity work that benefits from a model that can plan, act, and follow through without losing coherence over a long session.
The model supports a 991,000-token context window with up to 64,000 output tokens, alongside chat, function calling, image understanding, file input, and reasoning in a single configuration. That combination makes it well-suited for agent deployments where an assistant needs to read documents, inspect visuals, call external tools, and produce complete outputs across a long-running session - without needing to switch models or break the workflow. Blueprint-driven automations and Skillset-powered agents both benefit from the depth of context Qwen 3.7 Max can maintain.
At $1.25 per million input tokens and $3.75 per million output tokens, Qwen 3.7 Max is priced in the mid-tier premium range - competitive for a flagship model that covers reasoning, multimodal input, and tool use in a single call. Teams that need reliable agent-grade execution without the cost of the most expensive frontier models will find the tradeoff practical for production workloads.
Qwen 3.7 Max is available now in the ChatBotKit model picker and through the API.