Ollama v0.31.2: MLX small-batch matmul kernel, llama.cpp build 9840, CUDA updates
Ollama v0.31.2 shipped July 7 after two release candidates on July 6. Key changes: MLX updated with a new small-batch matmul kernel contributing to ~90% Gemma 4 throughput gains on Apple Silicon, llama.cpp updated to build 9840, CUDA and GPU compatibility improvements, UTF-8 file handling fixes, and unsupported ROCm device entries removed.
Why it matters
The MLX matmul kernel improvement produces measurable inference speedups for Apple Silicon users running popular models, the largest segment of the Ollama user base.
Importance: 2/5
Meaningful MLX throughput gain on Apple Silicon; regular maintenance release.