v0.40.0 v0.34.4
ollama/ollamaSep 25, 2026 → Sep 23, 20260 days between releases
Generating AI comparison
Analyzing what changed between v0.40.0 and v0.34.4. This page updates automatically — usually within a minute.
The Two Releases
v0.40.0
Sep 25, 2026· 27 days ago
Ollama now defaults to the MLX runtime on Apple Silicon, enabling native performance for supported models like Gemma4 and Qwen variants. This release extends support for decision models and embedding models to the MLX engine.
Full release notesv0.34.4
Sep 23, 2026· 29 days ago
This release focuses on performance improvements for Apple Silicon devices and thinking models, alongside critical bug fixes for macOS app responsiveness and model discovery. It also introduces enhanced image resolution handling for Gemma 4 and updates core libraries.
Full release notes