## What's Changed
* fix: improve max_tokens handling with conservative defaults by @Mingholy in https://github.com/QwenLM/qwen-code/pull/2438
* chore: bump version to 0.12.6 by @qwen-code-ci-bot in https://github.com/QwenLM/qwen-code/pull/2442
**Full Changelog**: https://github.com/QwenLM/qwen-code/compare/v0.12.5...v0.12.6