Ollama released version 0.40.7 on October 18, 2026. The update adds pplx‑decider, a 27‑billion‑parameter decision model from Perplexity that is fine‑tuned from Qwen3.8‑27B. It also fixes OpenAI‑compatible response handling, streaming tool calls, and remote model cancellation when the client disconnects. New API fields now report a model’s renderer, parser, and chat‑template source.
Why it matters
Developers can now use a larger decision model and rely on more stable streaming and remote model behavior in their Ollama‑based applications.
