Starstruk Logo MarkStarstruk

v0.32.1

knadh/listmonkv0.32.1Jul 16, 2026by github-actions[bot]

AI Summary

Focuses on improving Gemma 4's reasoning capabilities, fixing MLX model cache memory leaks, and enhancing agent context awareness.

Key Highlights

  • Improved tool calling and multi-turn reasoning for Gemma 4
  • Fixed MLX model cache leak that increased memory usage
  • Agent web search/fetch now prompts for authentication via ollama signin
  • Interactive agent now receives current working directory for better context
  • MLX text model loading now respects OLLAMA_LOAD_TIMEOUT

New Features

  • Enhanced Gemma 4 tool calling
  • Fixed MLX cache memory leak
  • Agent authentication prompts
  • Agent working directory context
  • MLX load timeout support

Full Release Notes

## What's Changed
- Improved Gemma 4 tool calling and multi-turn reasoning, including more reliable tool-response continuations
- Fixed a recurrent MLX model cache leak that could increase memory use across requests, and improved cache snapshot performance
- MLX text model loading now respects `OLLAMA_LOAD_TIMEOUT`
- Agent web search and fetch now tell users to run `ollama signin` when authentication is required
- The interactive agent now receives the current working directory for better project context
- Fixed `ollama launch` so choosing **Pick another model** for a deprecated model passed with `--model` opens the model picker
- Updated VS Code setup documentation for the official Ollama extension

**Full Changelog**: https://github.com/ollama/ollama/compare/v0.32.0...v0.32.1-rc0