Original release notes
What's Changed
- Support Laguna on Apple GPUs via the MLX engine
- Quantize draft-model output heads at the requested type when creating speculative-decoding drafts.
- Fixed Qwen3 MoE decoding for differently-quantized experts, plus faster packed gate/up projection (~4-9% on M5 Max).
Full Changelog: https://github.com/ollama/ollama/compare/v0.32.3...v0.32.4
