Ollama: Weekly Recap - MLX Model Expansion and Launch Polish
Ollama shipped 30 pull request actions with a focus on MLX model coverage and runner efficiency, plus tighter launch integrations. CUDA reliability fixes and OpenAI compatibility updates round out the week.
Duration: PT3M38S
Episode overview
This episode is a short developer briefing from Ollama.
It explains recent repository work in plain language.
- Show: Ollama
- Published: 2026-11-09T09:07:26Z
- Audio duration: PT3M38S
Transcript excerpt
This excerpt keeps the crawler page concise. Listen to the episode or use the RSS feed for the full update.
Hello, this is your Ollama briefing for November 2nd through November 9th.
This week: 30 pull request activity items, 22 additional commits this week.
The lead pattern is expansion paired with efficiency. The MLX runner added major new model families while cutting memory use and hardening accelerator startup, and the launch command cleaned up routing across third-party coding clients.
First, MLX model coverage. PR 19592 adds Ripple preview support to the MLX runner, including per-layer no-position encoding for draft models and fused attention and activation paths, with test models published for evaluation. PR 19231 honors explicit rotary scaling metadata in Qwen 3 and Qwen 3 point 5, applying…
Second, runner and engine reliability. PR 19521 rejects unsupported CUDA devices at startup using the architectures the build was compiled for, avoiding confusing load failures later. PR 19433 preloads bundled CUDA libraries in-process on Linux so the server does not pick up mismatched host copies. On Windows, PR…
Third, launch workflow and API compatibility. PR 19460 lets Pi auto-populate its chat model catalog. PR 19506 adds Kimi Code command line support through its config file. PRs 19587,…
Nearby episodes from Ollama
- Faster Concurrent Inference and Unified Desktop Launch
- MLX Goes Parallel
- VRAM Fixes and Launch Polish
- Stricter Model Imports and CLI Output Fixes
- Ripple on MLX and Validation Hardening
- MLX Memory Cuts and Tool-Call Fixes
- Launcher Return, API Fixes and Inference Reliability
- MLX Upgrades and OpenAI Compatibility Fixes