# Ollama v0.6.6 — Release & Upgrade Intelligence
- **Product:** Ollama (https://ollama.com)
- **Version:** v0.6.6
- **Released:** Apr 17, 2025
- **Impact Score:** 5/10
- **Breaking Changes:** Not verified
- **Canonical Analysis:** https://changelogs.us/ollama/updates/v0.6.6/

## Executive Summary (TL;DR)
Ollama v0.6.6. Source highlight: IBM Granite 3.3: 2B and 8B models with 128K context length that have been fine-tuned for improved reasoning and instruction-following capabilities.

## Improvements
- IBM Granite 3.3: 2B and 8B models with 128K context length that have been fine-tuned for improved reasoning and instruction-following capabilities.
- **New, faster model downloading**: `OLLAMA_EXPERIMENT=client2 ollama serve` will run Ollama using a new downloader with improved performance and reliability when running `ollama pull`. Please share feedback here!
- Improved performance of `ollama create` when importing models from Safetensors

## Bug Fixes
- Fixed memory leak issues when running Gemma 3, Mistral Small 3.1 and other models on Ollama
- Fixed certain out of memory issues from not reserving enough memory at startup
- Fix nondeterministic model unload order
- Fixed issue where model tensor data would be corrupted when importing models from Safetensors