Ollama 0.40.16 adds Ripple Preview, a 65B model for agentic coding

Ollama, the tool for running language models locally, released version 0.40.16 on 9 November.

A new coding model

The release adds Ripple Preview from Ricursive Intelligence to the model library. It is a 65-billion-parameter dense model built for agentic coding:

ollama run ripple-preview

Lower memory use and API fixes

• Qwen3.8 Flash Next uses about 27 GiB less memory on Apple silicon.
• The /v1/responses endpoint now respects the truncation setting and no longer truncates silently by default.
• /v1/models returns an empty list instead of null when no models are installed.
• ollama launch claude always routes Claude Code to the selected model, and ollama launch now explains when a cloud model has been retired.

Platforms

The release adds CUDA 13.5 support for Windows on ARM and updates the bundled llama.cpp.