Ollama 0.35 runs Nimble, Tev1 locally

Ollama positions version 0.35 as the first in a series of releases that support decision models, rather than a one-off addition. The models it can now run — Nimble, Tev1, and Tev1:0.8b — are built on the same idea as Jev, and the company frames their use around practical tasks such as ticket triaging, model routing, and content moderation.

The release is accompanied by an announced roadmap: future updates are planned to add MLX support for faster operation on Mac and to expand the number of supported models. That suggests Ollama intends to grow the decision-model line over time rather than stop at this first set.

For people already using Ollama to run AI locally, the significance hinges on how quickly those planned updates arrive and whether more decision models follow. The outcome depends on the roadmap being delivered, not just announced, and on whether these models prove useful for the triage, routing, and moderation tasks the company highlights.