Ollama Releases v0.40.7-rc0 with New Features and Changelog

What happened

Ollama has released version v0.40.7-rc0, which includes several new features and updates. The openai feature allows responses from non-thinking models, introduced by hoyyeva through a pull request tied to issue #19010. The llama-server update enables the system to infer the draft speculative type, implemented by dhiltgen via a pull request connected to issue #19064.

Additionally, the mlx feature adds support for PPLX Decider v1.2, developed by dhiltgen through a pull request related to issue #19072. The release notes mention a full changelog that allows users to compare changes between versions v0.40.6 and v0.40.7-rc0. This changelog is available for review.

Why it matters

These updates enhance the functionality of Ollama, making it more versatile for developers and users. The inclusion of the changelog allows for better tracking of changes and improvements, which is essential for maintaining and updating systems. The new features could impact how models are used and managed, improving efficiency and usability.

Key details

The release is part of the GitHub releases for Ollama, a project hosted on GitHub. The pull requests for the changes are associated with specific issues on the GitHub repository for Ollama. The version v0.40.7-rc0 is a release candidate for the Ollama project, and the changelog provides a detailed comparison between v0.40.6 and v0.40.7-rc0.

The mlx feature supports PPLX Decider v1.2, and the openai feature allows responses from non-thinking models. The llama-server update enables the system to infer the draft speculative type. These updates were implemented through pull requests tied to issues #19010, #19064, and #19072.