[MLX] Fix ValueError: Image features and image tokens do not match for Qwen3 VL Thinking models
Fix occasional UI crash when searching models
Build 1
Fixed streaming mode bug impacting tool-calling for Qwen 3 Coder in the /v1/chat/completions API (bug #1071).
[Vulkan] Fixed a bug where models were not being loaded onto iGPUs (requires runtime update) (bug #1048).
Added compatibility support for the developer role in the /v1/responses API endpoint. For now, developer messages are processed as system messages internally (bug #1064).
Resources
We are hiring! Check out our careers page for open roles.