Saturday, 21 November 2026, story 1 of 7
Ollama adds Thinking Machines' Inkling 2 Mini for Apple silicon Macs
Ollama says the 26B model (5.4B active) fits in about 20GB and scores 79.2% on SWE-Bench Verified. PC support is promised, not shipped.

What happened
Ollama's v0.40.17 release, published 10 November, adds Inkling 2 Mini from Thinking Machines. Run it with ollama run inkling-2-mini. Ollama describes a 26B-parameter model with 5.4B active, which it says fits on a Mac at about 20GB. It handles text and images, tool calls and thinking, and Ollama gives a score of 79.2% on SWE-Bench Verified.
Support is Apple silicon only in this release. Ollama says PC support will follow in an upcoming release, with no date. The same release lets Ollama on Apple silicon serve multiple requests at once. Hugging Face has also published a post welcoming the model.
What it means
The practical news is the footprint. A model Ollama says fits in about 20GB, with 5.4B parameters active per token, puts a coding-focused model within reach of a well-equipped Mac. Serving multiple requests at once on Apple silicon helps if several tools or agents share one local model.
Treat the 79.2% as a reported figure. The source gives no test setup, and no independent results appear in it. Teams on Windows or other PCs cannot run the model through Ollama yet, and the source gives no date. It also states no license terms, which matters before anything ships in a product. Worth watching: independent coding results and the PC release.
What to do: If you have a Mac with memory to spare, try ollama run inkling-2-mini on your own code tasks before trusting the 79.2% figure; check the license first.
Sources
- v0.40.17, Ollama releases
- Welcome Inkling 2 Mini, Cutting-edge Performance for your Local Setup, Hugging Face
Filed under Open source. Published .
"What it means" is analysis, not reporting. Found an error? Write to daily@vibitely.com.