Ollama v0.31.1 Released with Faster Gemma 4 on Apple Silicon

Core summary

One sentence to understand this update

Ollama has released v0.31.1, which significantly boosts Gemma 4 performance on Apple Silicon, achieving nearly 90% faster token generation via multi-token prediction.

Impact & opportunity

What this could mean

Apple Silicon developers can now expect much faster Gemma 4 model inference with Ollama, improving local development efficiency.