ollaya-dev/ollaya v0.14.0 is a release that reflects side-by-side measurement results across multiple models and Ollama versions, decoder tags, and content on model lifecycle. The repository is a project that runs public judgment models locally, downloading and serving Laya, decider, NLI and GLiClass behind a TypeSafe-compatible API. The source of this change is the GitHub release record published on October 19, 2026. The release summary points readers to the measurements and the numeric data together.
The purpose of this project is to run judgment models locally and provide them through an API. So the key thing to look at in this release is not a description that concludes which model is better, but material that lets you view measurement results across multiple models and Ollama versions side by side. Decoder tags and model lifecycle are covered in the same release. However, the provided information does not present measured values or measurement conditions for each comparison target, so performance differences or suitable environments cannot be determined from this content alone.
To check the numbers and detailed results, you need to look at the v0.14.0 release record directly. The given summary only supports that measurement data exists and that the record asks you to check it alongside. Which model is faster or more accurate, and which Ollama version is better, cannot be said here. So this update is best read as a record with comparison evidence added, and performance conclusions should be drawn only after checking the published values and conditions.