v0.22.0

Ollama has released version 0.22.0, bringing two notable new models to its local AI runtime platform. The update includes NVIDIA's Nemotron 3 Omni, a multimodal model that can process text, images, and other data types, and Poolside's Laguna XS.2, the company's first open-weight coding model designed specifically for code generation and completion tasks. This release expands Ollama's model library, which now supports over 170,000 stars on GitHub.

The addition of these models gives developers and AI enthusiasts more options for running advanced AI locally. Nemotron 3 Omni offers multimodal reasoning capabilities that were previously only available through cloud APIs, while Laguna XS.2 provides a specialized coding model that can be used offline for tasks like code review, debugging, and generation. Both models are available for download and use through Ollama's standard interface, requiring no additional configuration beyond the standard setup.