Ollama has released version 0.32.11, bringing several notable enhancements for developers working with local large language models. The update includes DeepSeek Harness, a tool designed to streamline model evaluation and benchmarking, which is particularly useful for teams fine-tuning or comparing models. Additionally, Muse Code integration offers a coding assistant directly within the Ollama ecosystem, potentially simplifying development workflows. The most significant addition is Web Search support via the Responses API, enabling models to access up-to-date information from the web, a feature that was previously limited to cloud-based services. This move aligns with the growing demand for hybrid local-cloud AI solutions. For developers, these features reduce the friction of building context-aware applications and improve the overall utility of local LLM deployments. While the update is incremental, it reflects Ollama's commitment to keeping pace with the evolving needs of the AI developer community.
Ollama's latest version introduces DeepSeek Harness for improved model evaluation, Muse Code integration for coding assistance, and Web Search support through the Responses API. These additions expand Ollama's utility for developers building local AI applications.