ai tool
Ollama: security advisories, deprecations and releases
Runs open-weight language models locally.
- Items, 90 days
- 30
- Critical
- 0
- Last critical
- none
- This week
- 5
What your digest would have held this week
5 items from the last 7 days. Zero sends only the ones that touch your stack.
Security 1
- HIGH Google Cloud Blog · 2 Oct 2026
2,076 AI-stack vulnerabilities disclosed
GTIG tracked 2,076 AI-related CVEs from January 2025 to August 2026, concentrated in orchestration frameworks, inference engines, and web portals.
Releases 29
- INFO NEW THIS WEEK Ollama releases · 3 Nov 2026
Ollama v0.40.14 release candidate
This release candidate fixes build deadlock, improves OpenAI response handling, MLX top-k sorting, ChatGPT retry logic, and other minor fixes.
- INFO NEW THIS WEEK Ollama releases · 29 Oct 2026
Ollama v0.40.13 released
Ollama 0.40.13 adds support for Liquid AI's LFM3 family on Linux and Windows, including text-to-speech and omni models.
- INFO NEW THIS WEEK Ollama releases · 29 Oct 2026
Ollama v0.40.13-rc0 released
Release candidate adds LFM3 GGUF fixes and updates llama.cpp to b11963.
- INFO NEW THIS WEEK Ollama releases · 27 Oct 2026
Ollama v0.40.12 adds LFM3 and d1 models
Ollama v0.40.12 adds support for LFM3 and d1 models via the MLX backend.
- LOW NEW THIS WEEK Ollama releases · 27 Oct 2026
Ollama 0.40.12 performance and compatibility fixes
Faster Gemma 4/Flint/Laguna prompt processing on Apple silicon, DFlash2 speculative decoding on MLX, improved small-image detail for Gemma 4, OpenCode v2 support, MLX and llama.cpp updates.
- INFO Ollama releases · 27 Oct 2026
Ollama v0.40.12-rc0 release
Release candidate adds MLX performance gains for Gemma/Flint/Laguna, DFlash2 support, and a llama.cpp bump.
- INFO Ollama releases · 24 Oct 2026
Ollama v0.40.11 released
Adds real-time text-to-speech on macOS via ollama run and OpenAI-compatible endpoint, plus d3 decision models (2B-27B) for classification and scoring.
- INFO Ollama releases · 24 Oct 2026
Ollama v0.40.11-rc0 released
Minor release fixes function tags, context handling, vllm-d3 state, llama.cpp bump, and adds Phonon voice catalog.
- INFO Ollama releases · 23 Oct 2026
Ollama v0.40.10 adds NVIDIA Nemotron 3.5 Bolt E4B
New release adds a small NVIDIA vision-language model with 128K context and speculative decoding for faster local use.
- INFO Ollama releases · 23 Oct 2026
Ollama v0.40.9-rc3 released
New release adds speech generation on Apple Silicon via Phonon, d3 decision models, Flint on Linux/Windows, plus bug fixes for streaming and context handling.
- INFO Ollama releases · 21 Oct 2026
Ollama v0.40.8 adds Flint model support
Ollama now supports Flint models by preserving numeric types in prompts, assembling expert banks during safetensors import, and adding learned sinks with device causal masks.
- INFO Ollama releases · 21 Oct 2026
Ollama v0.40.8-rc0 adds Flint coding model
Ollama 0.40.8-rc0 introduces Flint, a 40B-parameter Apache-2.0 model for repository-level coding tasks on macOS.
- MEDIUM Ollama releases · 21 Oct 2026
Ollama v0.40.7 released
Minor update improves Responses API compatibility, fixes streaming tool-call ordering, adds model renderer/parser metadata, and updates llama.cpp and MLX.
- INFO Ollama releases · 18 Oct 2026
Ollama v0.40.7-rc0 released
Release candidate adds OpenAI Responses effort support for non-thinking models, llama-server draft speculative type inference, and PPLX Decider v1.2 on MLX.
- INFO Ollama releases · 17 Oct 2026
Ollama v0.40.6-rc1 released
Fixes Windows orphaned runners, adds Q2_0 tensor support, launches Clef in Ollama Cloud, improves --ask mode.
- INFO Ollama releases · 16 Oct 2026
Ollama v0.40.5 released
Ollama 0.40.5 adds NVIDIA Nemotron 3.5 Super/Bolt vision-language models, cloud decision-model support, OpenAI-compatible responses fixes, and updated llama.cpp/MLX.
- INFO Ollama releases · 14 Oct 2026
Ollama v0.40.5-rc0 released
Release candidate adds CUDA port of Metal kernels, custom tool-call support, docs updates, and MLX bump.
- INFO Ollama releases · 14 Oct 2026
Ollama v0.40.4 released
Adds decision models to CLI, improves MLX runner on Apple silicon, and boosts performance across NVIDIA and multimodal embeddings.
- INFO Ollama releases · 13 Oct 2026
Ollama v0.40.4-rc1 released
This release candidate updates llama.cpp, reuses verified manifests, compacts MLX state, adds Codex model picker, and bumps MLX versions.
- INFO Ollama releases · 12 Oct 2026
Ollama v0.40.4-rc0 released
Update brings llama.cpp to b11521 for new models, manifest reuse for faster inference, MLX state fixes, and Codex model picker updates.
- LOW Ollama releases · 11 Oct 2026
Ollama v0.40.3 released
Fixes spurious warnings when launching Claude and Codex, plus paused background model upgrade for better request performance.
- INFO Ollama releases · 8 Oct 2026
Ollama v0.40.2 released
Ollama v0.40.2 includes merged pull request 18855 adding Claude max context tokens support.
- INFO Ollama releases · 8 Oct 2026
Ollama v0.40.1 released
Minor release fixes Windows symlink and memory issues, removes CLI onboarding step, and improves documentation links.
- LOW Ollama releases · 6 Oct 2026
Ollama v0.40.0 fixes embed route 413 vs 400 routing
Ollama v0.40.0 corrects the routing of 413 vs 400 errors for embed requests, improving error handling consistency.
- MEDIUM Ollama releases · 5 Oct 2026
Ollama v0.40.0-rc3: MLX runtime default on Apple Silicon
Ollama now runs MLX-supported models automatically on Apple Silicon, expanding supported model list.
- LOW Ollama releases · 3 Oct 2026
Ollama v0.34.4 released
Fixes intermittent model-not-found errors, improves structured outputs on thinking models, speeds up Qwen 3.8 on Apple Silicon and picks optimal image resolutions for Gemma 4.
- INFO Ollama releases · 3 Oct 2026
Ollama v0.40.0-rc0 released
On Apple Silicon, models supported by MLX will now run on the MLX runtime by default.
- INFO Ollama releases · 3 Oct 2026
Ollama v0.35.0 adds decision-model API
New /v1/systemone endpoint returns choices, probabilities, and scores instead of text for classification tasks.
- INFO Ollama releases · 3 Oct 2026
Ollama v0.35.1-rc2 release notes
Ollama 0.35.1-rc2 adds support for ten web searches per response, bumps MLX and llama.cpp, and enables explicit model capabilities.
Only hear about Ollama when it touches your stack
Tell Zero what you run. You get one email, and only on the days something above applies to you.