AI-ML
Ollama v0.32.6
RESUMEN
What's Changed - Qwen3.5 is faster on Apple GPUs: the MLX engine now uses the model's MTP head for speculative decoding automatically - `/v1/chat/completions` streaming now matches OpenAI's wire format: `role` only on the first chunk, `finish_reason` on its own chunk, and usage in a sepa
Descripción Detallada
What's Changed - Qwen3.5 is faster on Apple GPUs: the MLX engine now uses the model's MTP head for speculative decoding automatically - `/v1/chat/completions` streaming now matches OpenAI's wire format: `role` only on the first chunk, `finishreason` on its own chunk, and usage in a separate chunk with `streamoptions.includeusage`. - Truncated OpenAI responses now report `finishreason: "length"` instead of `"tool_calls"`. - `ollama run kimi-k3` now offers `kimi-k3:cloud` for cloud-only models that publish no default tag, instead of failing. - TUI fixes: pipe-delimited prose no longer renders as a table, Enter accepts the highlighted `@` file completion, and `/prompt` scrolling is no longer laggy. - Experimental image generation has been temporarily removed. Continue using 0.32.5 for image generation support - Updated the MLX and llama.cpp engines. Full Changelog:
Explicación con IA
Genera un resumen en lenguaje claro de los cambios de este release.
Releases Relacionados
AI-ML
Ollama v0.32.5
## What's Changed * Fixed an MLX Metal bug that could reduce output quality for NVFP4 models, particularly Laguna. **Full Changelog**: https://github.com/ollama/ollama/compare/v0.32.4...v0.32.5
AI-ML
Ollama v0.32.4
## 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
AI-ML