Qwen3.6-Plus Officially Launched with 1 Million Context Window, Compatible with Claude Code and OpenClaw

According to monitoring by 1M AI News, Alibaba’s Tongyi Qianwen team has released Qwen3.6-Plus, the next-generation flagship model following the Qwen3.5 series launched in February this year. The API is now available for use via Alibaba Cloud’s Bailing service. The core upgrade focuses on enhancing Agent programming capabilities while strengthening multimodal perception and reasoning, with a default support for a 1 million token context window. In programming, Qwen3.6-Plus excels in benchmarks for code repair, terminal operations, and automated task execution, achieving optimal results in various long-range planning and tool invocation benchmarks. Front-end development capabilities have also been enhanced, allowing it to handle complex projects such as 3D scenes and games. In terms of multimodal capabilities, the model shows improvements in tasks like complex document understanding, video reasoning, and visual programming, supporting scenarios such as generating code from interface screenshots and restoring front-end pages based on design drafts. The API introduces a new parameter, preserve_thinking, which retains reasoning content from previous rounds in multi-turn dialogues. The official recommendation is to use this for Agent tasks, as maintaining complete reasoning context helps improve decision consistency and reduce token consumption from repetitive reasoning. Alibaba Cloud Bailing supports both OpenAI-compatible and Anthropic-compatible protocols, the latter allowing direct integration with Claude Code. It is also compatible with programming tools such as OpenClaw, Qwen Code, Kilo Code, Cline, and OpenCode. Users of Qwen Code can log in via OAuth to receive 1,000 free calls daily. The Qianwen team has stated that their recent focus will shift towards the overall release of the Qwen3.6 series, with plans to open-source smaller model versions in the future.