Ollama: System One Decisions Launch
Ollama merged a new System One scoring API for local structured decisions, with MLX support and docs following. Reliability fixes for stalled downloads, chat truncation, and expanded web search round out the release activity.
Duration: PT2M14S
Episode overview
This episode is a short developer briefing from Ollama.
It explains recent repository work in plain language.
- Show: Ollama
- Published: 2026-09-29T13:01:29Z
- Audio duration: PT2M14S
Transcript excerpt
This excerpt keeps the crawler page concise. Listen to the episode or use the RSS feed for the full update.
Good morning, it's Tuesday, September 29th, 2026, and this is your Ollama developer briefing.
The headline: Ollama is adding a new kind of local model — one built not to chat, but to decide.
That work centers on pull request 18606, merged yesterday, which adds a System One scoring A P I. You send in context and named questions, and using local Nimble and Tev models, you get back a structured answer: a choice with probabilities, a yes-or-no probability, or an expected score across an ordered rubric.…
That one merge is part of a larger push. An open follow-up, 18701, adds M L X support for those decision models and wires the capability into the Model file, while 18702 adds the Decision guide and A P I reference with examples and limits. For developers, this means expect a documented, local-first decision endpoint…
The second theme is reliability when work gets long. Pull request 18625 fixes a stall in safety tensor downloads: previously the watchdog could detect a stalled pull but couldn't cancel it, and retries were unbounded with no backoff. Now the request owns cancellation, and stall retries get a bounded budget with…
Finally, a product shift: pull request 18700 would make desktop…
Nearby episodes from Ollama
- Silent Sampling Overrides Fixed
- Weekly Recap - Single-Pass Structured Outputs and MLX Speedups
- Tool-Call Reliability Fixes
- Desktop Fixes and API Correctness
- Mac Polish, MLX Reliability, and Compatibility Fixes
- GPU Efficiency Fix and MLX Hardening
- Single-Pass Structured Outputs and MLX Speedups
- Agent Reliability Fixes and Offline Model Transfer