← November 4, 2026 briefing

Ollama v0.40.14 applies model output limits when launching Claude Code and respects sampler defaults in OpenAI-compatible API; v0.40.15-rc0 adds MiMo V2.6

Announced November 3, 2026

Verified (fact)

In v0.40.14, 'ollama launch claude' applies per-model output token limits so cloud models no longer hit the 128k default cap. The OpenAI-compatible endpoint now uses the model's built-in defaults when temperature and top_p are not sent, and decoding of cloud web search responses was fixed. The subsequent v0.40.15-rc0 added support for the MiMo V2.6 model.

Why it matters (interpretation)

It reduces output truncation and sampling mismatches that users hit when connecting local and cloud models to coding agents.

Sources

Story thread

  1. 2026-11-04 Ollama v0.40.14 applies model output limits when launching Claude Code and respects sampler defaults in OpenAI-compatible API; v0.40.15-rc0 adds MiMo V2.6
  2. 2026-11-07 Ollama v0.40.15 released: adds support for Xiaomi's MiMo V2.6 (9B) and the 'AutoTrust' decision model family

More about Ollama