# Ollama changelog > Get up and running with large language models locally. - Vendor: Ollama - Category: AI - License: MIT - Language: Go - Official site: https://ollama.com - Tracked by: What's New (https://whatsnew.fyi/product/ollama) - Harvested from: GitHub (ollama/ollama) - Entries below: 25 (newest first) What's New is an index, not a publisher: every entry below links to the vendor's own release notes, which are the authoritative source. Entries are labelled where they are hand-curated sample data, pre-releases, or drawn from a secondary source such as a developer blog. Reuse: release facts (products, versions, dates, links, CVE ids and flags) are open data under CC BY 4.0, from the API at https://whatsnew.fyi/api/v1 or the nightly dataset at https://github.com/whatsnew-fyi/dataset. The summaries, labels and curation are © What's New: quote freely with attribution and a link back; wholesale republication of them is not permitted — terms: https://whatsnew.fyi/terms. The vendors' own release notes remain their publishers'. ## Releases ### v0.40.5 - Date: 2026-10-14 - Version: v0.40.5 - Original notes: https://github.com/ollama/ollama/releases/tag/v0.40.5 - Permalink: https://whatsnew.fyi/product/ollama/releases/v0.40.5-rc0 - **added** — NVIDIA Nemotron 3.5 Super vision-language model, runnable with ollama run nemotron-3.5-super - **added** — Nemotron 3.5 Bolt, a 74B-parameter distilled sibling of Super tuned to fit a 24 GB NVIDIA GPU, with speculative decoding via its included drafter - **added** — Decision models can be run through ollama.com cloud using the clef:cloud and clef-flash:cloud tags without pulling them locally - **fixed** — Continuing an OpenAI-compatible Responses conversation that includes custom tool calls no longer fails with an error - **changed** — Cloud access error messages now clearly explain sign-in problems, payment issues in the desktop app, and upstream access errors from the ChatGPT gateway - **changed** — Updated llama.cpp, adding Nemotron 3.5 support, and MLX ##### NVIDIA Nemotron 3.5 Super [NVIDIA Nemotron 3.5 Super](https://ollama.com/library/nemotron-3.5-super) is a vision-language model for coding, reasoning, tool use, and document understanding. It pairs a 120B-parameter hybrid Mamba-2/Transformer mixture-of-experts backbone (12B active per token) with a C-RADIO vision encoder, and supports tool calling, an on/off thinking mode, and context up to 1M tokens. ``` ollama run nemotron-3.5-super ``` A compact sibling, [Nemotron 3.5 Bolt](https://ollama.com/library/nemotron-3.5-bolt), distills Super down to 74B parameters (9B active) tuned to fit a 24 GB NVIDIA GPU for everyday agentic tasks, with speculative decoding via its included drafter. ##### What's Changed * Decision models are now available on cloud — use `clef:cloud` and `clef-flash:cloud` to run decision models through ollama.com cloud without pulling them locally * Continuing an OpenAI-compatible Responses conversation that includes custom tool calls no longer fails with an error * Clearer error messages when cloud access fails: sign-in problems, payment issues in the desktop app, and upstream access errors from the ChatGPT gateway now explain what went wrong * Updated llama.cpp (adding Nemotron 3.5 support) and MLX ##### New Contributors * @willietran made their first contribution in https://github.com/ollama/ollama/pull/18986 **Full Changelog**: https://github.com/ollama/ollama/compare/v0.40.4...v0.40.5 ### v0.40.4 - Date: 2026-10-12 - Version: v0.40.4 - Original notes: https://github.com/ollama/ollama/releases/tag/v0.40.4 - Permalink: https://whatsnew.fyi/product/ollama/releases/v0.40.4-rc0 - **added** — Decision models can now be used directly from the CLI with `ollama run`, for yes/no questions, text classification, and rubric scoring - **added** — Decision models are now supported by the MLX runner on Apple silicon - **changed** — `ollama launch` can now be used with usage credits without upgrading to a paid plan - **changed** — `ollama launch codex` now includes locally installed Ollama models in its model picker, and `/model` within the Codex CLI can be used to swap models - **changed** — Inference performance improved by reusing verified manifests and already loaded models - **added** — Support added for multimodal embeddings through llama-server - **fixed** — Loading multimodal embedding models with small context windows is fixed - **changed** — Gemma 4 prefill performance improved on NVIDIA GPUs with the MLX runner - **changed** — MLX runner now clips long Laya states instead of rejecting them - **changed** — Speculative decoding reliability improved in the MLX runner, including recurrent state handling after rollback - **changed** — llama.cpp updated with native decision model support - **changed** — README and documentation updated with clearer privacy and cloud-services wording ##### Decision models in the CLI Decision models can now be used directly from the terminal with `ollama run` — ask yes/no questions, classify text into labels, or score against a rubric: `ollama run nimble --ask "Is a refund being requested?" < ticket.txt` ## What's Changed * Kev decision models are now supported by the MLX runner on Apple silicon * `ollama launch` can now be used with usage credits without upgrading to a paid plan * `ollama launch codex` now includes locally installed Ollama models in its model picker. Use `/model` within the Codex CLI to swap models. * Improved inference performance by reusing verified manifests and already loaded models * Added support for multimodal embeddings through llama-server * Fixed loading multimodal embedding models with small context windows * Improved Gemma 4 prefill performance on NVIDIA GPUs with the MLX runner * MLX runner now clips long Laya states instead of rejecting them * Improved speculative decoding reliability in the MLX runner, including recurrent state handling after rollback * Updated llama.cpp with native decision model support * Updated README and documentation with clearer privacy and cloud-services wording **Full Changelog**: https://github.com/ollama/ollama/compare/v0.40.3...v0.40.4 ### v0.40.3 - Date: 2026-10-11 - Version: v0.40.3 - Original notes: https://github.com/ollama/ollama/releases/tag/v0.40.3 - Permalink: https://whatsnew.fyi/product/ollama/releases/v0.40.3-rc0 - **fixed** — `ollama launch claude` no longer shows a warning about Claude connectors - **fixed** — `ollama launch codex` no longer warns about service tiers - **changed** — Paused background model upgrade for embedding and other request performance ##### What's Changed * `ollama launch claude` no longer shows a warning about Claude connectors by @hoyyeva in https://github.com/ollama/ollama/pull/18878 * `ollama launch codex` no longer warns about service tiers by @hoyyeva in https://github.com/ollama/ollama/pull/18879 * Paused background model upgrade for embedding and other request performance, by @jmorganca in https://github.com/ollama/ollama/pull/18908 **Full Changelog**: https://github.com/ollama/ollama/compare/v0.40.2...v0.40.3 ### v0.40.2 - Date: 2026-10-08 - Version: v0.40.2 - Original notes: https://github.com/ollama/ollama/releases/tag/v0.40.2 - Permalink: https://whatsnew.fyi/product/ollama/releases/v0.40.2-rc0 - **changed** — Models downloaded with earlier versions are upgraded in the background the first time they are run, for better performance and compatibility when running on llama.cpp - **changed** — Original copies of upgraded models are kept as backups on disk to allow safe downgrading, and a future release will remove these backups automatically - **fixed** — `ollama list` no longer shows duplicate entries for upgraded models - **changed** — `ollama launch claude` uses the model's full context length ##### Model upgrades Models downloaded with earlier versions of Ollama are upgraded in the background the first time you run them, for better performance and compatibility when running on llama.cpp. To make downgrading safe, Ollama keeps the original copy as a backup, so upgraded models are temporarily kept on disk. A future release will remove these backups automatically. To remove backed up models now (requires `jq`): ```shell curl -s localhost:11434/api/tags | jq -r '.models[].name' | while read -r m; do curl -s localhost:11434/api/show -d "{\"model\":\"$m\"}" | jq -r \ 'select(any(.manifests[]?; .runner == "llamacpp")) | .manifests[] | select(.runner == "ggml") | .digest' done | sort -u | xargs -n1 ollama rm ``` This only deletes the backups. If you later downgrade to a version older than 0.40, you'll need to re-pull those models. **Other changes** * `ollama list` no longer shows duplicate entries for upgraded models by @dhiltgen in https://github.com/ollama/ollama/pull/18874 * `ollama launch claude` uses the model's full context length by @jberg5 in https://github.com/ollama/ollama/pull/18855 * README: add oxi to community integrations by @maziluiosif in https://github.com/ollama/ollama/pull/18739 ##### New Contributors * @maziluiosif made their first contribution in https://github.com/ollama/ollama/pull/18739 **Full Changelog**: https://github.com/ollama/ollama/compare/v0.40.1...v0.40.2-rc0 ### v0.40.1 - Date: 2026-10-07 - Version: v0.40.1 - Original notes: https://github.com/ollama/ollama/releases/tag/v0.40.1 - Permalink: https://whatsnew.fyi/product/ollama/releases/v0.40.1-rc0 - **added** — Server proxy cloud usage and balance APIs - **fixed** — Fix CLEF head reads past 2GiB on Windows - **removed** — Remove account step from CLI onboarding - **changed** — Avoid symlinks on Windows in manifest handling - **fixed** — Fix dead links in README community integrations list - **fixed** — Fix broken download links in app README - **changed** — Drop carried metal residency patch now that it is upstream in MLX ##### What's Changed * server: proxy cloud usage and balance APIs by @drifkin in https://github.com/ollama/ollama/pull/18829 * llama: fix clef head reads past 2GiB on windows by @Gigrise in https://github.com/ollama/ollama/pull/18777 * cmd: remove account step from CLI onboarding by @hoyyeva in https://github.com/ollama/ollama/pull/18826 * manifest: avoid symlinks on Windows by @dhiltgen in https://github.com/ollama/ollama/pull/18852 * docs: fix 6 dead links in README community integrations list by @aniketkrs in https://github.com/ollama/ollama/pull/18814 * docs: fix broken download links in app README by @chenlichao in https://github.com/ollama/ollama/pull/18233 * mlx: drop carried metal residency patch now that it is upstream by @dhiltgen in https://github.com/ollama/ollama/pull/18854 ##### New Contributors * @Gigrise made their first contribution in https://github.com/ollama/ollama/pull/18777 * @aniketkrs made their first contribution in https://github.com/ollama/ollama/pull/18814 * @chenlichao made their first contribution in https://github.com/ollama/ollama/pull/18233 **Full Changelog**: https://github.com/ollama/ollama/compare/v0.40.0...v0.40.1-rc0 ### v0.35.1 - Date: 2026-09-29 - Version: v0.35.1 - Original notes: https://github.com/ollama/ollama/releases/tag/v0.35.1 - Permalink: https://whatsnew.fyi/product/ollama/releases/v0.35.1-rc0 - **added** — Support for Clef and Clef Flash decision models through /v1/systemone endpoint - **added** — Multimodal support for decision models allowing images to be included alongside text state in requests - **changed** — Models using web search can now perform up to ten searches per response, increased from three - **added** — Modelfiles now support CAPABILITY declarations to explicitly declare what a model can do - **changed** — ollama show and the model list now report only decision as the capability for decision models - **changed** — Updated llama.cpp and the MLX engine ##### Clef decision models Ollama now supports [Clef](https://ollama.com/library/clef) and [Clef Flash](https://ollama.com/library/clef), Cloudflare's new open-source decision models, through `/v1/systemone`. Clef (27B) and Clef Flash (9B) are multimodal: requests can now include images alongside the text state, shared by all questions and scored jointly with it. ```sh curl http://localhost:11434/v1/systemone -d '{ "model": "clef-flash", "state": "The user took this screenshot.", "images": [""], "questions": { "has_ollama": {"type": "noul", "instructions": "Does this image contain Ollama?"} } }' ``` ``` { "model": "clef-flash", "answers": { "has_ollama": { "type": "noul", "noul": 0.958 } }, "usage": { "input_tokens": 548, "output_tokens": 0 } } ``` ##### What's Changed * Models using web search can now perform up to ten searches per response, up from three * Modelfiles now support `CAPABILITY` declarations, so model creators can explicitly declare what a model can do. Declarations are preserved when creating from GGUF or safetensors, through model inheritance, and on Modelfile export * `ollama show` and the model list now report only `decision` as the capability for decision models, so clients no longer offer them for general chat, tools, or thinking * Updated llama.cpp and the MLX engine **Full Changelog**: https://github.com/ollama/ollama/compare/v0.35.0...v0.35.1 ### v0.35.0 - Date: 2026-09-28 - Version: v0.35.0 - Original notes: https://github.com/ollama/ollama/releases/tag/v0.35.0 - Permalink: https://whatsnew.fyi/product/ollama/releases/v0.35.0-rc1 - **added** — Support decision models through /v1/systemone endpoint based on TypeSafe's Jev API - **added** — Decision models available: Nimble from Bespoke Labs and Tev1 from Together AI - **added** — Support three question types for decision models: choice, noul, and score - **changed** — Settings now opens without waiting for model discovery - **fixed** — Fixed the macOS update menu and icon not reflecting an available update at startup - **fixed** — Fixed stalled MLX model downloads hanging indefinitely - **changed** — Requests containing the deprecated typical_p parameter now log a warning instead of failing ##### Decision models Ollama now supports decision models through `/v1/systemone`, based on [TypeSafe’s Jev API](https://typesafe.ai). Decision models return choices, probabilities, and scores instead of text. Use them for tasks such as ticket triage, model routing, and content classification. Available models: - [**Nimble**](https://ollama.com/library/nimble) from Bespoke Labs - [**Tev1**](https://ollama.com/library/tev1) from Together AI ```sh ollama pull nimble ``` Send context and one or more questions: ```sh curl http://localhost:11434/v1/systemone \ -H 'Content-Type: application/json' \ -d '{ "model": "nimble", "state": "Our checkout has returned 500 errors since 9am.", "questions": { "label": { "type": "choice", "instructions": "Which label fits this ticket?", "criteria": { "billing": "Payments and refunds", "bug": "Software errors", "account": "Login and account access" } } } }' ``` Example response: ```json { "model": "nimble", "answers": { "label": { "type": "choice", "choice": "bug", "probabilities": { "billing": 0.0125, "bug": 0.9781, "account": 0.0093 }, "confidence": 0.8906 } }, "usage": { "input_tokens": 174, "output_tokens": 1 } } ``` The API supports three question types: - `choice`: Select an option and return probabilities for each. - `noul`: Return the probability that a condition is true. - `score`: Return a score across an ordered set of criteria ##### What's Changed - Settings now opens without waiting for model discovery. - Fixed the macOS update menu and icon not reflecting an available update at startup. - Fixed stalled MLX model downloads hanging indefinitely. - Requests containing the deprecated `typical_p` parameter now log a warning instead of failing. **Full Changelog:** https://github.com/ollama/ollama/compare/v0.34.4...v0.35.0 ### v0.40.0 - Date: 2026-09-25 - Version: v0.40.0 - Original notes: https://github.com/ollama/ollama/releases/tag/v0.40.0 - Permalink: https://whatsnew.fyi/product/ollama/releases/v0.40.0-rc0 - **changed** — Model architectures supported by the MLX runtime automatically run on MLX on Apple Silicon devices by default - **added** — MLX support for decision models including Nimble, tev1, clef, and clef-flash - **added** — MLX support for embedding model embeddinggemma-2 ##### What's Changed **Models run on MLX on Apple Silicon by default** In this release, on Apple Silicon devices, model architectures supported by the MLX runtime will automatically run on MLX. ``` ollama pull qwen3.8 ollama run qwen3.8 ``` Additional models include [gemma4](https://ollama.com/library/gemma4), [qwen3.6](https://ollama.com/library/qwen3.6) and [qwen3.5](https://ollama.com/library/qwen3.5) Decision models are now available on MLX as well: [Nimble](https://ollama.com/library/nimble) [tev1](https://ollama.com/library/tev1) [clef](https://ollama.com/library/clef) [clef-flash](https://ollama.com/library/clef-flash) MLX now has support for an embedding model: [embeddinggemma-2](https://ollama.com/library/embeddinggemma-2) We will continue testing and enabling additional models. **Full Changelog**: https://github.com/ollama/ollama/compare/v0.35.1...v0.40.0 ### v0.34.4 - Date: 2026-09-23 - Version: v0.34.4 - Original notes: https://github.com/ollama/ollama/releases/tag/v0.34.4 - Permalink: https://whatsnew.fyi/product/ollama/releases/v0.34.4-rc0 - **changed** — Structured outputs on thinking models now apply in a single pass, making them faster and more reliable - **fixed** — Fixed intermittent "model not found" errors with a large local library - **fixed** — Fixed the macOS app becoming unresponsive when checking if ChatGPT or Codex is running - **changed** — Qwen 3.8 prompt processing is faster on Apple Silicon - **changed** — Gemma 4 on Apple Silicon now picks the best image resolution per image, keeping more detail in high-resolution images - **changed** — Updated llama.cpp, MLX, and XGrammar ##### What's Changed - Structured outputs on thinking models now apply in a single pass, making them faster and more reliable. - Fixed intermittent "model not found" errors with a large local library - Fixed the macOS app becoming unresponsive when checking if ChatGPT or Codex is running. - Qwen 3.8 prompt processing is faster on Apple Silicon. - Gemma 4 on Apple Silicon now picks the best image resolution per image, keeping more detail in high-resolution images. - Updated llama.cpp, MLX, and XGrammar. **Full Changelog**: https://github.com/ollama/ollama/compare/v0.34.3...v0.34.4 ### v0.34.3 - Date: 2026-09-19 - Version: v0.34.3 - Original notes: https://github.com/ollama/ollama/releases/tag/v0.34.3 - Permalink: https://whatsnew.fyi/product/ollama/releases/v0.34.3-rc1 - **added** — GET /api/show endpoint now advertises each model's thinking controls and default values - **added** — Nemotron H vision models are now supported on Apple Silicon with MLX - **changed** — Ollama's macOS app will no longer reopen windows you've closed when activating it - **fixed** — Fix for model pulls from HuggingFace ##### What's Changed `GET /api/show` now advertises each model's thinking controls and default: Available in the CLI with: ``` ollama show gemma4 ``` ``` thinking levels false, true default true ``` Available in the API with: ```sh curl http://localhost:11434/api/show -d '{"model": "glm-5.3-flash:cloud"}' ``` ```json { "thinking": { "values": ["low", "high", "max"], "default": "max" } } ``` Also available on ollama.com directly for cloud models. * **Nemotron H** vision models are now supported on Apple Silicon with MLX * Ollama's macOS app will now no longer reopen windows you've closed when activating it * Fix for model pulls from HuggingFace **Full Changelog**: https://github.com/ollama/ollama/compare/v0.34.2...v0.34.3 ### v0.34.2 - Date: 2026-09-15 - Version: v0.34.2 - Original notes: https://github.com/ollama/ollama/releases/tag/v0.34.2 - Permalink: https://whatsnew.fyi/product/ollama/releases/v0.34.2-rc0 - **added** — First-run setup when running ollama with options to sign in or continue locally, with setup completion shared with the desktop app on macOS and Windows - **added** — ollama://apps protocol handler to open the desktop app's Apps page directly on macOS and Windows - **fixed** — Excessive memory growth during long generations with MLX speculative decoding - **changed** — Updated llama.cpp ##### What's Changed - Added first-run setup when running `ollama`, with options to sign in or continue locally. Setup completion is shared with the desktop app on macOS and Windows. - Added `ollama://apps` to open the desktop app’s Apps page directly on macOS and Windows. - Fixed excessive memory growth during long generations with MLX speculative decoding. - Updated llama.cpp. **Full Changelog**: https://github.com/ollama/ollama/compare/v0.34.1...v0.34.2 ### v0.34.1 - Date: 2026-09-14 - Version: v0.34.1 - Original notes: https://github.com/ollama/ollama/releases/tag/v0.34.1 - Permalink: https://whatsnew.fyi/product/ollama/releases/v0.34.1-rc1 - **changed** — MLX safetensors ollama create is no longer experimental - **changed** — GGUF model creation now requires using llama.cpp tooling for safetensor conversion and quantization - **changed** — Improved MLX memory handling on Apple Silicon - **changed** — Runaway repeat token detection now requires 100 repeat tokens for reduced false positives - **changed** — /api/tags is much faster on large model libraries, now with consistent model capabilities reporting - **deprecated** — typical_p can no longer be set when creating new models, though existing GGUF models retain support ##### What's Changed * MLX safetensors `ollama create` no longer experimental. GGUF model creation now requires using llama.cpp tooling for safetensor conversion and quantization. * Improved MLX memory handling on Apple Silicon * Runaway repeat token detection now requires 100 repeat tokens for reduced false positives (e.g. OCR) * `/api/tags` is much faster on large model libraries (3.1 s → 294 ms cold in testing), and model capabilities are now reported consistently. * Deprecated `typical_p`: it can no longer be set when creating new models, existing GGUF models retain support. * MLX and llama.cpp updates **Full Changelog**: https://github.com/ollama/ollama/compare/v0.34.0...v0.34.1-rc1 ### v0.34.0 - Date: 2026-09-05 - Version: v0.34.0 - Original notes: https://github.com/ollama/ollama/releases/tag/v0.34.0 - Permalink: https://whatsnew.fyi/product/ollama/releases/v0.34.0-rc1 - **added** — Support for using Ollama models directly in ChatGPT Desktop with setup available from the Ollama app on macOS - **changed** — Improved structured output performance on Apple Silicon - **added** — Support for OpenAI-compatible client tool search - **added** — Support for response compaction ##### Use Ollama models in ChatGPT Desktop Ollama models can now be used directly in ChatGPT Desktop, so you can keep your existing workflow while running open models. Setup is available from the Ollama app on MacOS. CleanShot 2026-09-08 at 11 04 07 AM@2x This release also improves structured output performance on Apple Silicon, adds support for OpenAI-compatible client tool search and response compaction. **Full Changelog**: https://github.com/ollama/ollama/compare/v0.33.3...v0.34.0 ### v0.33.3 - Date: 2026-09-02 - Version: v0.33.3 - Original notes: https://github.com/ollama/ollama/releases/tag/v0.33.3 - Permalink: https://whatsnew.fyi/product/ollama/releases/v0.33.3-rc0 - **added** — gemma4 now supports images and audio on MLX engine - **added** — Report cached prompt tokens - **changed** — Honor GGUF model defined default parameters - **changed** — MLX, MLX-C, and llama.cpp updated ##### What's Changed * gemma4 now supports images and audio on MLX engine * Report cached prompt tokens * Honor GGUF model defined default parameters * MLX, MLX-C, llama.cpp update ##### New Contributors * @marcelpetrick made their first contribution in https://github.com/ollama/ollama/pull/17579 **Full Changelog**: https://github.com/ollama/ollama/compare/v0.33.2...v0.33.3 ### v0.33.2 - Date: 2026-08-27 - Version: v0.33.2 - Original notes: https://github.com/ollama/ollama/releases/tag/v0.33.2 - Permalink: https://whatsnew.fyi/product/ollama/releases/v0.33.2-rc1 - **fixed** — Ollama's app now follows the system appearance again, restoring dark mode support - **fixed** — Fixed the macOS app to properly hand off to an already-running instance instead of starting a second one - **fixed** — The Claude Desktop proxy no longer interrupts in-flight requests when the model catalog updates ##### What's Changed * Ollama's app now follows the system appearance again, restoring dark mode support * Fixed the macOS app to properly hand off to an already-running instance instead of starting a second one * The Claude Desktop proxy no longer interrupts in-flight requests when the model catalog updates **Full Changelog**: https://github.com/ollama/ollama/compare/v0.33.1...v0.33.2 ### v0.33.1 - Date: 2026-08-26 - Version: v0.33.1 - Original notes: https://github.com/ollama/ollama/releases/tag/v0.33.1 - Permalink: https://whatsnew.fyi/product/ollama/releases/v0.33.1-rc1 - **added** — Add Qwen3.8 Flash Next support for MLX - **changed** — Make external compat patches idempotent in cmake - **changed** — Update MLX and llama.cpp - **added** — Add structured output support to mlxrunner - **fixed** — Avoid Metal GPU timeouts when loading models from slow storage in mlxrunner ##### What's Changed * MLX: Qwen3.8 Flash Next support * cmake: make external compat patches idempotent * MLX and llama.cpp update * mlxrunner: add structured output support * mlxrunner: avoid Metal GPU timeouts when loading models from slow storage ##### New Contributors * @pd95 made their first contribution in https://github.com/ollama/ollama/pull/17948 **Full Changelog**: https://github.com/ollama/ollama/compare/v0.33.0...v0.33.1 ### v0.33.0 - Date: 2026-08-21 - Version: v0.33.0 - Original notes: https://github.com/ollama/ollama/releases/tag/v0.33.0 - Permalink: https://whatsnew.fyi/product/ollama/releases/v0.33.0-rc2 - **added** — Claude Desktop can now be configured to work with Ollama as a third-party gateway provider - **fixed** — Fixed a hang where agent clients that cancel long prefills - **changed** — Prefill restore points are now trustworthy by construction so retries resume where they stopped instead of restarting from scratch - **fixed** — Resumed prefills no longer record restore points that fail to cover what they claim, preventing reprocessing from zero on models with recurrent layers - **fixed** — Disabled Claude Code's tokens left token-countdown system message which broke the KV cache on every request - **changed** — DeepSeek Harness launcher now falls back to npx when the global npm install fails, with Windows command-shim support - **changed** — Updated MLX dependency - **fixed** — Fixed broken default packaging caused by macOS-specific assumptions affecting Linux and Windows builds ##### What's Changed ###### Claude Desktop Developers can now easily configure Claude Desktop to seamlessly work with Ollama as a third-party gateway provider. image ###### Improved caching * Fixed a hang where agent clients that cancel long prefills * Prefill restore points are now trustworthy by construction: a cancelled prefill keeps every restore point it crossed, so retries resume where they stopped instead of restarting from scratch * Resumed prefills no longer record restore points that fail to cover what they claim; on models with recurrent layers this previously forced a request matching 46k of 47k tokens to reprocess from zero * Disabled Claude Code's "tokens left" token-countdown system message, which Ollama moved to the front of the prompt and broke the KV cache on every request ###### Other improvements * DeepSeek Harness launcher now falls back to `npx` when the global npm install fails, with Windows command-shim support * MLX dependency update (#17886) * Fixed broken default packaging caused by macOS-specific assumptions affecting Linux/Windows builds **Full Changelog**: https://github.com/ollama/ollama/compare/v0.32.15...v0.33.0 ### v0.32.15 - Date: 2026-08-19 - Version: v0.32.15 - Original notes: https://github.com/ollama/ollama/releases/tag/v0.32.15 - Permalink: https://whatsnew.fyi/product/ollama/releases/v0.32.15-rc1 - **added** — New desktop onboarding flow on first launch - **changed** — Cache resolved model metadata between requests to reduce time-to-first-token - **fixed** — Fix a bug where chat and generate could wedge after a mid-stream parser error - **changed** — Normalize Qwen 3.8 system messages so non-leading system messages are handled consistently - **changed** — Update MLX and llama.cpp dependencies ##### What's Changed * New desktop onboarding flow on first launch * Caches resolved model metadata between requests, cutting time-to-first-token by roughly half (TTFT dropped from ~995 ms to ~524 ms in benchmarks) * Fixes a bug where chat and generate could wedge after a mid-stream parser error * **Qwen 3.8** system messages are now normalized so non-leading system messages are handled consistently * MLX and llama.cpp dependency updates ##### New Contributors * @gaugarg-nv made their first contribution in https://github.com/ollama/ollama/pull/17752 **Full Changelog**: https://github.com/ollama/ollama/compare/v0.32.14...v0.32.15 ### v0.32.14 - Date: 2026-08-15 - Version: v0.32.14 - Original notes: https://github.com/ollama/ollama/releases/tag/v0.32.14 - Permalink: https://whatsnew.fyi/product/ollama/releases/v0.32.14-rc0 - **fixed** — Transcode WebP images for llama-server - **fixed** — Tolerate non-leading system messages in Qwen renderer ##### What's Changed * llm: transcode WebP images for llama-server * renderers/qwen: tolerate non-leading system messages **Full Changelog**: https://github.com/ollama/ollama/compare/v0.32.13...v0.32.14 ### v0.32.13 - Date: 2026-08-14 - Version: v0.32.13 - Original notes: https://github.com/ollama/ollama/releases/tag/v0.32.13 - Permalink: https://whatsnew.fyi/product/ollama/releases/v0.32.13 - **added** — Support developer instructions for qwen2.5-7b model ##### What's Changed * qwen3.8: support developer instructions **Full Changelog**: https://github.com/ollama/ollama/compare/v0.32.12...v0.32.13 ### v0.32.12 - Date: 2026-08-14 - Version: v0.32.12 - Original notes: https://github.com/ollama/ollama/releases/tag/v0.32.12 - Permalink: https://whatsnew.fyi/product/ollama/releases/v0.32.12 - **added** — Support for Qwen 3.8 27B model - **added** — Optimized qwen3.8:27b-mlx variant for Apple Silicon devices ollama and qwen capybara ##### Qwen 3.8 27B This release adds the support of [Qwen 3.8 27B](https://ollama.com/library/qwen3.8). Qwen3.8 delivers substantial gains across coding, professional work, research, and long-horizon agentic tasks. ``` ollama run qwen3.8:27b ``` For Apple Silicon devices, Ollama has in particular optimized for maximum performance and output quality suitable for repeated tasks and coding agents. ``` ollama run qwen3.8:27b-mlx ``` ### v0.32.11 - Date: 2026-08-14 - Version: v0.32.11 - Original notes: https://github.com/ollama/ollama/releases/tag/v0.32.11 - Permalink: https://whatsnew.fyi/product/ollama/releases/v0.32.11 - **added** — ollama launch dsh now supports DeepSeek Harness, DeepSeek's open-source agent harness - **added** — ollama launch muse now supports Muse Code, Meta's agentic coding CLI - **added** — The OpenAI-compatible Responses API now supports web search - **changed** — Muse Glimmer template updates ##### What's Changed * `ollama launch dsh` now supports DeepSeek Harness, [DeepSeek's open-source agent harness](https://deepseek.com/harness/en/) * `ollama launch muse` now supports [Muse Code](https://developer.meta.com/ai/products/muse-code/), Meta's agentic coding CLI * The OpenAI-compatible Responses API now supports web search * Muse Glimmer template updates **Full Changelog**: https://github.com/ollama/ollama/compare/v0.32.10...v0.32.11 ### v0.32.10 - Date: 2026-08-12 - Version: v0.32.10 - Original notes: https://github.com/ollama/ollama/releases/tag/v0.32.10 - Permalink: https://whatsnew.fyi/product/ollama/releases/v0.32.10-rc1 - Labels: Pre-release - **changed** — Models that don't set a repeat_penalty now default to 1.0 (off) instead of 1.1, matching other engines and speeding up speculative decoding - **changed** — Faster prefill on NVFP4 MLX models with a global scale, about 7–8% on Qwen3.6 and Muse Glimmer - **fixed** — Fixed blob verification being skipped when an OCI manifest's config and layer share a digest ##### What's Changed - Models that don't set a `repeat_penalty` now default to 1.0 (off) instead of 1.1, matching other engines and speeding up speculative decoding; set a per-model parameter if an older model repeats itself. - Faster prefill on NVFP4 MLX models with a global scale, about 7–8% on Qwen3.6 and Muse Glimmer. - Fixed blob verification being skipped when an OCI manifest's config and layer share a digest. ##### New Contributors * @vigneshakaviki made their first contribution in https://github.com/ollama/ollama/pull/15504 **Full Changelog**: https://github.com/ollama/ollama/compare/v0.32.8...v0.32.10-rc1 ### v0.32.9 - Date: 2026-08-11 - Version: v0.32.9 - Original notes: https://github.com/ollama/ollama/releases/tag/v0.32.9 - Permalink: https://whatsnew.fyi/product/ollama/releases/v0.32.9 - **added** — Added the Nemotron 3 architecture - **fixed** — Handle boundary condition in Muse Glimmer function calling parser ##### NVIDIA Nemotron 3.5 Lightning [NVIDIA Nemotron 3.5 Lightning](https://ollama.com/library/nemotron-3.5-lightning) is an open 30B mixture-of-experts (MoE) model with 3B active parameters built for that execution layer of always-on agents. It is designed for harnesses like OpenClaw and Hermes Agent – all supported by the [NVIDIA NemoClaw](https://github.com/NVIDIA/NemoClaw) open source security and management stack for running always-on AI agents. ``` ollama run nemotron-3.5-lightning ``` ##### What's Changed * Added the Nemotron 3 architecture * Handle boundary condition in Muse Glimmer function calling parser **Full Changelog**: https://github.com/ollama/ollama/compare/v0.32.8...v0.32.9 ### v0.32.8 - Date: 2026-08-10 - Version: v0.32.8 - Original notes: https://github.com/ollama/ollama/releases/tag/v0.32.8 - Permalink: https://whatsnew.fyi/product/ollama/releases/v0.32.8-rc0 - **added** — Add Muse Glimmer support for NVIDIA, AMD, and additional platforms ##### Muse Glimmer Muse Glimmer is now available on all platforms. Muse Glimmer can power coding agent applications such as Claude Code, Codex, Pi and more, as well as long-running personal assistants such as OpenClaw and Hermes. Ollama's MLX engine provides state-of-the-art performance on Apple Silicon for this model, with support for DFlash and image input as of Ollama 0.32.7. To download and run Muse Glimmer locally: ```bash ollama run muse-glimmer ``` To run Muse Glimmer with Claude Code, [download Ollama](https://ollama.com/download) and run: ```bash ollama launch claude --model muse-glimmer ``` For a lighter-weight coding agent, try [Pi](https://pi.dev): ```bash ollama launch pi --model muse-glimmer ``` For personal assistant frameworks such as OpenClaw and Hermes, use: ```bash ollama launch openclaw --model muse-glimmer ``` ```bash ollama launch hermes --model muse-glimmer ``` ##### What's Changed * Add Muse Glimmer support for NVIDIA, AMD, and additional platforms **Full Changelog**: https://github.com/ollama/ollama/compare/v0.32.7...v0.32.8