Skip to content
AIBRAINWORX / OpenSourceCoder.comLOCAL-FIRST. BUILT IN THE OPEN.
Inference & Quantization / Developer & Agent Tools / Inference & Hardware / Models & Weights

Ollama v0.40.17-rc1 adds MLX batch speculative verification

The November 16 release candidate updates Ollama's Apple Silicon mlxrunner while keeping recent MiMo V2.6 and AutoTrust support.

Listen to this articleBrowser Voice
Ready
GitHub release social card for ollama/ollama v0.40.17-rc1
AI-generated conceptual illustration by AiBrainWorX; not a photograph or product screenshot. · Ollama project icon; identification only, no endorsement.
Event / context: Ollama v0.40.17-rc1 was published November 16, 2026.. Source check: October 28, 2026.

Ollama v0.40.17-rc1 introduces batch speculative verification in its mlxrunner backend, targeting faster token acceptance on Apple Silicon systems.

Pre-release status

As a release candidate (-rc1), the build is intended for testing ahead of stable promotion. Recent v0.40.x updates also added support for Xiaomi MiMo V2.6, AutoTrust decision models, and Kimi Code CLI integration in ollama launch.

EXPLORE THE SOURCE

GitHub repositories

Original reporting; primary project releases reviewed November 16, 2026. No independent hands-on benchmark claimed.

Follow your curiosity.

Pass the signal.