Ollama released v0.40.15 on November 3, 2026. The release adds two model families and several launcher and serving improvements. Read the release notes.
New models
- mimo-v2.6: Xiaomi MiMo's 9B model built for code and agent workflows:
ollama run mimo-v2.6. - autotrust-jev: AutoTrust decision models that read text and return choices, yes/no probabilities and ratings; the 35b and 27b-vl sizes also read images. Supported on Apple silicon in this release.
Other changes
ollama launchnow supports the Kimi Code CLI (successor to kimi-cli) and offers to migrate existing kimi-cli setups; it also fills in Pi's model list automatically, and Claude Desktop's model picker shows Ollama's actual model names.- Recurrent and vision GGUF models can now serve more than one request at a time.
- Qwen3 and Qwen3.5 models gain long-context YaRN scaling.
- Running
ollamawith no arguments opens a welcome launcher.