This release adds 4 notable features for engineering teams evaluating rollout.
✓ No known CVEs patched in this version
Topics
+8 more
Summary
AI summaryFixed ollama launch provider selection bug.
Changes in this release
| Type | Severity | Summary | CVE |
|---|---|---|---|
| Feature | Low |
MLX runner now creates snapshots during prompt processing and speculative decoding for improved reliability MLX runner now creates snapshots during prompt processing and speculative decoding for improved reliability Source: llm_adapter@2026-06-12 Confidence: high |
— |
| Feature | Low |
Improves recurrent model support with per-boundary states from the gated-delta kernels Improves recurrent model support with per-boundary states from the gated-delta kernels Source: llm_adapter@2026-06-12 Confidence: high |
— |
| Performance | Low |
Improves prompt caching by decoupling it from context shift for better KV cache reuse Improves prompt caching by decoupling it from context shift for better KV cache reuse Source: llm_adapter@2026-06-12 Confidence: high |
— |
| Performance | Low |
Increases stability of MLX inference with hardened linear and embedding layers Increases stability of MLX inference with hardened linear and embedding layers Source: llm_adapter@2026-06-12 Confidence: high |
— |
| Bugfix | Medium |
Fixes `ollama launch` selecting the wrong provider in some cases Fixes `ollama launch` selecting the wrong provider in some cases Source: llm_adapter@2026-06-12 Confidence: high |
— |
Full changelog
What's Changed
- Fixed
ollama launchselecting the wrong provider in some cases - Improved prompt caching by decoupling it from context shift for better KV cache reuse
- More stable MLX inference with hardened linear and embedding layers
- MLX runner now creates snapshots during prompt processing and speculative decoding for improved reliability
- Improved recurrent model support with per-boundary states from the gated-delta kernels
Full Changelog: https://github.com/ollama/ollama/compare/v0.30.7...v0.30.8
Weekly OSS security release digest.
The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.
No spam, unsubscribe anytime.
Share this release
About ollama
Get up and running with Kimi-K2.5, GLM-5, MiniMax, DeepSeek, gpt-oss, Qwen, Gemma and other models.
Beta — feedback welcome: [email protected]