Ollama v0.30.0 Release: Direct llama.cpp Support and GGUF Compatibility

Core summary

One sentence to understand this update

Ollama v0.30.0 introduces a new architecture that directly supports llama.cpp, replacing GGML, and enables compatibility with the GGUF file format, while also utilizing MLX for accelerated inference on Apple devices.

Impact & opportunity

What this could mean

Developers can benefit from improved performance and wider model compatibility on Ollama, especially on Apple hardware, simplifying local LLM deployment and experimentation.