{"id":"src_6zK8-hxSfdq7TiQRvyd5M","slug":"ollama","name":"Ollama","type":"github","url":"https://github.com/ollama/ollama","orgId":"org_JwNQHJi6geHoZZWC1PTlV","productId":null,"productSlug":null,"org":{"id":"org_JwNQHJi6geHoZZWC1PTlV","slug":"ollama","name":"Ollama"},"isPrimary":true,"isHidden":false,"discovery":"curated","metadata":"{\"wellKnownSweptAt\":\"2026-06-24T06:00:01.224Z\"}","notice":null,"kind":null,"stars":174847,"starsFetchedAt":"2026-06-24T16:05:14.445Z","releaseCount":104,"releasesLast30Days":5,"avgReleasesPerWeek":2.1,"latestVersion":"v0.30.10","latestDate":"2026-06-17T16:22:02.000Z","changelogUrl":null,"hasChangelogFile":false,"lastFetchedAt":"2026-06-24T16:05:14.445Z","lastPolledAt":"2026-06-24T16:05:04.951Z","trackingSince":"2025-03-14T02:58:48.000Z","releases":[{"id":"rel_Trl_LRx41HNk1EWHO6D-t","version":"v0.30.10","type":"feature","title":"v0.30.10","summary":"## What's Changed\r\n\r\n* Command A and North family models now run on Apple Silicon with the MLX engine\r\n* Updated the underlying llama.cpp engine to bu...","titleGenerated":null,"titleShort":null,"content":"## What's Changed\r\n\r\n* Command A and North family models now run on Apple Silicon with the MLX engine\r\n* Updated the underlying llama.cpp engine to build 9672\r\n* Fixed build artifacts for MLX\r\n\r\n**Full Changelog**: https://github.com/ollama/ollama/compare/v0.30.9...v0.30.10\r\n","publishedAt":"2026-06-17T16:22:02.000Z","fetchedAt":"2026-06-18T05:04:40.992Z","url":"https://github.com/ollama/ollama/releases/tag/v0.30.10","media":[],"coverageCount":0},{"id":"rel_SPCWOOqfsvf_hrLP-lMR2","version":"v0.30.9","type":"feature","title":"v0.30.9","summary":"## What's Changed\r\n* Support for Cohere2Moe architecture\r\n* Fixed LFM2 parser/render for cases where thinking was not emitted\r\n* Fixed issue where `ol...","titleGenerated":null,"titleShort":null,"content":"## What's Changed\r\n* Support for Cohere2Moe architecture\r\n* Fixed LFM2 parser/render for cases where thinking was not emitted\r\n* Fixed issue where `ollama launch claude` and other coding agent or assistant use cases would only output one token\r\n* Ollama will now return an error if a single message is larger than the current context window\r\n\r\n**Full Changelog**: https://github.com/ollama/ollama/compare/v0.30.8...v0.30.9-rc1","publishedAt":"2026-06-15T19:55:07.000Z","fetchedAt":"2026-06-17T04:05:21.375Z","url":"https://github.com/ollama/ollama/releases/tag/v0.30.9","media":[],"coverageCount":0},{"id":"rel_1zk4i9zVT9Dlo4UUVgj_Y","version":"v0.30.5","type":"feature","title":"v0.30.5","summary":"## What's Changed\r\n* Fix gemma4:12b floating point exception crash\r\n* integrations: hermes windows install by @BruceMacD in https://github.com/ollama/...","titleGenerated":null,"titleShort":null,"content":"## What's Changed\r\n* Fix gemma4:12b floating point exception crash\r\n* integrations: hermes windows install by @BruceMacD in https://github.com/ollama/ollama/pull/16487\r\n\r\n**Full Changelog**: https://github.com/ollama/ollama/compare/v0.30.4...v0.30.5","publishedAt":"2026-06-04T17:00:37.000Z","fetchedAt":"2026-06-04T21:05:12.922Z","url":"https://github.com/ollama/ollama/releases/tag/v0.30.5","media":[],"coverageCount":0},{"id":"rel__Yas6Nx6o7uybvkb_wLvC","version":"v0.30.3","type":"feature","title":"v0.30.3","summary":"## What's Changed\r\n* models: add support for gemma4-12b by @pdevine in https://github.com/ollama/ollama/pull/16457\r\n\r\n\r\n**Full Changelog**: https://gi...","titleGenerated":null,"titleShort":null,"content":"## What's Changed\r\n* models: add support for gemma4-12b by @pdevine in https://github.com/ollama/ollama/pull/16457\r\n\r\n\r\n**Full Changelog**: https://github.com/ollama/ollama/compare/v0.30.2...v0.30.3","publishedAt":"2026-06-03T16:35:43.000Z","fetchedAt":"2026-06-03T20:05:08.914Z","url":"https://github.com/ollama/ollama/releases/tag/v0.30.3","media":[],"coverageCount":0},{"id":"rel_R5KTBVhtdT7CmVAwtMOyV","version":"v0.30.2","type":"feature","title":"v0.30.2","summary":"## What's Changed\r\n* feat(launch): show and auto-install Cline CLI by @hoyyeva in https://github.com/ollama/ollama/pull/16402\r\n* log template details ...","titleGenerated":null,"titleShort":null,"content":"## What's Changed\r\n* feat(launch): show and auto-install Cline CLI by @hoyyeva in https://github.com/ollama/ollama/pull/16402\r\n* log template details to aid troubleshooting by @dhiltgen in https://github.com/ollama/ollama/pull/16403\r\n* cmd/launch: add Qwen code integration by @hoyyeva in https://github.com/ollama/ollama/pull/15900\r\n* launch: fix opencode local model limits by @dhiltgen in https://github.com/ollama/ollama/pull/16425\r\n* llm: include cached prompt tokens in llama-server counts by @dhiltgen in https://github.com/ollama/ollama/pull/16428\r\n* Harden app markdown URL handling by @dhiltgen in https://github.com/ollama/ollama/pull/16380\r\n* discover: allow Radeon 8060S iGPU by default by @dhiltgen in https://github.com/ollama/ollama/pull/16429\r\n* llm: detect llama-server load stalls from output by @dhiltgen in https://github.com/ollama/ollama/pull/16427\r\n* More harden app markdown URL handling by @dhiltgen in https://github.com/ollama/ollama/pull/16436\r\n* llama.cpp version update by @dhiltgen in https://github.com/ollama/ollama/pull/16426\r\n* launch: isolate Codex launch configuration by @ParthSareen in https://github.com/ollama/ollama/pull/16437\r\n* llama: add laguna (poolside) arch via a llama.cpp patch under llama/c… by @dhiltgen in https://github.com/ollama/ollama/pull/16396\r\n* docs: configure hermes desktop app by @BruceMacD in https://github.com/ollama/ollama/pull/16440\r\n* llm: ignore llama-server SSE ping comments by @dhiltgen in https://github.com/ollama/ollama/pull/16443\r\n* fix laguna patch build breakage by @dhiltgen in https://github.com/ollama/ollama/pull/16445\r\n\r\n\r\n**Full Changelog**: https://github.com/ollama/ollama/compare/v0.30.0...v0.30.2-rc0","publishedAt":"2026-06-03T00:37:18.000Z","fetchedAt":"2026-06-03T05:04:53.927Z","url":"https://github.com/ollama/ollama/releases/tag/v0.30.2","media":[],"coverageCount":0},{"id":"rel_I9gdDhTCHgQinfISBbaTi","version":"v0.24.0","type":"feature","title":"v0.24.0","summary":"## Codex App\r\n\r\nOllama 0.24 includes support for the Codex App, OpenAI's desktop experience for working on Codex threads in parallel with built-in wor...","titleGenerated":null,"titleShort":null,"content":"## Codex App\r\n\r\nOllama 0.24 includes support for the Codex App, OpenAI's desktop experience for working on Codex threads in parallel with built-in worktree support and git functionality.\r\n\r\n```bash\r\nollama launch codex-app\r\n```\r\n\r\n<img width=\"2088\" height=\"1404\" alt=\"CleanShot 2026-05-14 at 15 04 18@2x\" src=\"https://github.com/user-attachments/assets/53bd7997-19fd-4809-b8f2-b6ed284369c9\" />\r\n\r\n\r\n### Built-in browser\r\nCodex can load local servers and sites in its built-in browser, enabling you to directly annotate on the page to request changes.\r\n\r\n<img width=\"1073\" height=\"668\" alt=\"codex-annotate copy\" src=\"https://github.com/user-attachments/assets/c9b762b3-83f2-47f1-8f28-d9eebc1bf5e0\" />\r\n\r\n\r\n### Review mode\r\nReview code inside the app, leave comments, and iterate without leaving your workspace.\r\n\r\n<img width=\"1137\" height=\"696\" alt=\"codex-comments copy 2\" src=\"https://github.com/user-attachments/assets/56316d33-59ed-4f24-aaa7-a7c0310014c4\" />\r\n\r\n### Choosing a model\r\n\r\nFor difficult coding and agentic tasks:\r\n\r\n- **kimi-k2.6** (with vision support)\r\n- **glm-5.1**\r\n\r\nFor local use without an Ollama Cloud subscription:\r\n\r\n- **nemotron-3-super**\r\n- **gemma4:31b**\r\n- **qwen3.6** \r\n\r\n### Restore anytime\r\n\r\nTo restore the previous configuration of Codex App, run:\r\n\r\n```bash\r\nollama launch codex-app --restore\r\n```\r\n\r\n## What's Changed\r\n\r\n* Reworked the MLX sampler for improved generation quality on Apple Silicon\r\n\r\n**Full Changelog**: https://github.com/ollama/ollama/compare/v0.23.0...v0.24.0\r\n","publishedAt":"2026-05-14T02:24:24.000Z","fetchedAt":"2026-05-26T11:36:05.501Z","url":"https://github.com/ollama/ollama/releases/tag/v0.24.0","media":[{"type":"image","url":"https://github.com/user-attachments/assets/53bd7997-19fd-4809-b8f2-b6ed284369c9","r2Key":"releases/6fdf179293a60a46cec74c03a8eb0a2efff7f058cfc96000bc77860304ecf8c1.png","r2Url":"https://media.releases.sh/releases/6fdf179293a60a46cec74c03a8eb0a2efff7f058cfc96000bc77860304ecf8c1.png"},{"type":"image","url":"https://github.com/user-attachments/assets/c9b762b3-83f2-47f1-8f28-d9eebc1bf5e0","r2Key":"releases/0f91bf7773c500ba224f861bee8a0784b4b4a9ae06e5d3b414cbee71422c8459.png","r2Url":"https://media.releases.sh/releases/0f91bf7773c500ba224f861bee8a0784b4b4a9ae06e5d3b414cbee71422c8459.png"},{"type":"image","url":"https://github.com/user-attachments/assets/56316d33-59ed-4f24-aaa7-a7c0310014c4","r2Key":"releases/f84f167fa890963884b09f27865654894a90bc547484a96ecd61db3b92779d4c.png","r2Url":"https://media.releases.sh/releases/f84f167fa890963884b09f27865654894a90bc547484a96ecd61db3b92779d4c.png"}],"coverageCount":0},{"id":"rel_PCC2hNG5hwjk8AlLSj8dZ","version":"v0.23.4","type":"feature","title":"v0.23.4","summary":"## What's Changed\r\n* `ollama launch opencode` now supports vision models with image inputs\r\n* Fixed formatting of Claude tool results when using local...","titleGenerated":null,"titleShort":null,"content":"## What's Changed\r\n* `ollama launch opencode` now supports vision models with image inputs\r\n* Fixed formatting of Claude tool results when using local image paths\r\n\r\n**Full Changelog**: https://github.com/ollama/ollama/compare/v0.23.3...v0.23.4","publishedAt":"2026-05-13T20:40:22.000Z","fetchedAt":"2026-05-26T11:36:05.501Z","url":"https://github.com/ollama/ollama/releases/tag/v0.23.4","media":[],"coverageCount":0},{"id":"rel_HMJXSp4SsH8_bZaYw5f9u","version":"v0.30.0","type":"feature","title":"v0.30.0","summary":"Ollama 0.30 is now available, with improved compatibility and performance using [llama.cpp](https://github.com/ggml-org/llama.cpp). This augments the ...","titleGenerated":null,"titleShort":null,"content":"Ollama 0.30 is now available, with improved compatibility and performance using [llama.cpp](https://github.com/ggml-org/llama.cpp). This augments the MLX engine on Apple Silicon, bringing support to a wider range of hardware.\r\n\r\nThis release brings support for a wider range of models, including GGUF-based models from Hugging Face and your own fine-tuned models along with faster performance on NVIDIA hardware.\r\n\r\n## Known issues:\r\n\r\n* `laguna-xs.2` is not yet supported on Windows/Linux.\r\n* `llama3.2-vision` is not yet supported\r\n* `nomic-embed-text` now converts inputs to lowercase per the model card where prior Ollama versions incorrectly preserved mixed case\r\n","publishedAt":"2026-05-13T14:32:54.000Z","fetchedAt":"2026-06-02T04:04:53.051Z","url":"https://github.com/ollama/ollama/releases/tag/v0.30.0","media":[],"coverageCount":0},{"id":"rel_J_1BtjAxlxBjTVdGE-jL_","version":"v0.23.3","type":"feature","title":"v0.23.3","summary":"## What's Changed\r\n* mlx: refined model push behavior by @dhiltgen in https://github.com/ollama/ollama/pull/15431\r\n* test: integration test hardening ...","titleGenerated":null,"titleShort":null,"content":"## What's Changed\r\n* mlx: refined model push behavior by @dhiltgen in https://github.com/ollama/ollama/pull/15431\r\n* test: integration test hardening by @dhiltgen in https://github.com/ollama/ollama/pull/13532\r\n* app: harden update flows by @dhiltgen in https://github.com/ollama/ollama/pull/16100\r\n* mlx: update the imagegen runner for mlx thread affinity by @pdevine in https://github.com/ollama/ollama/pull/16096\r\n* mlx: avoid status timeout during inference by @dhiltgen in https://github.com/ollama/ollama/pull/16086\r\n* mlx: fix macOS 26 target leakage in v3 metallib by @dhiltgen in https://github.com/ollama/ollama/pull/16053\r\n\r\n\r\n**Full Changelog**: https://github.com/ollama/ollama/compare/v0.23.2...v0.23.3","publishedAt":"2026-05-12T03:48:08.000Z","fetchedAt":"2026-05-26T11:36:05.501Z","url":"https://github.com/ollama/ollama/releases/tag/v0.23.3","media":[],"coverageCount":0},{"id":"rel_Fcy4w54yuFLtKaJgD6USd","version":"v0.23.2","type":"feature","title":"v0.23.2","summary":"## What's Changed\r\n\r\n* `ollama launch` no longer includes Claude Desktop due to the third-party integration being limited to Anthropic models. \r\n* Use...","titleGenerated":null,"titleShort":null,"content":"## What's Changed\r\n\r\n* `ollama launch` no longer includes Claude Desktop due to the third-party integration being limited to Anthropic models. \r\n* Use `ollama launch claude-desktop --restore` to restore Claude Desktop to its normal state.\r\n* `/api/show` responses are now cached, improving median latency by **~6.7x** which will increase load speed for integrations like VS Code.\r\n* Improved backup workflow when managing launch integrations\r\n* Cleaner image generation layout in the MLX runner\r\n\r\n**Full Changelog**: https://github.com/ollama/ollama/compare/v0.23.1...v0.23.2","publishedAt":"2026-05-07T20:23:10.000Z","fetchedAt":"2026-05-26T11:36:05.501Z","url":"https://github.com/ollama/ollama/releases/tag/v0.23.2","media":[],"coverageCount":0},{"id":"rel_-WP3Ue2Ohc94NQY6Snf2c","version":"v0.23.1","type":"feature","title":"v0.23.1","summary":"## Gemma 4 MTP (Multi-token Processing) for the MLX runner\r\nGemma 4 MTP speculative decoding is now supported on Macs. This can give over a 2x speed i...","titleGenerated":null,"titleShort":null,"content":"## Gemma 4 MTP (Multi-token Processing) for the MLX runner\r\nGemma 4 MTP speculative decoding is now supported on Macs. This can give over a 2x speed increase for the Gemma 4 31B model on coding tasks.\r\n\r\n```\r\nollama run gemma4:31b-coding-mtp-bf16\r\n```\r\n\r\n## What's Changed\r\n* Update MLX and MLX-C with threading fixes by @dhiltgen in https://github.com/ollama/ollama/pull/15845\r\n* go: bump to 1.26 by @ParthSareen in https://github.com/ollama/ollama/pull/15904\r\n* Add Gemma 4 MTP speculative decoding by @pdevine in https://github.com/ollama/ollama/pull/15980\r\n\r\n**Full Changelog**: https://github.com/ollama/ollama/compare/v0.23.0...v0.23.1","publishedAt":"2026-05-05T17:13:31.000Z","fetchedAt":"2026-05-26T11:36:05.999Z","url":"https://github.com/ollama/ollama/releases/tag/v0.23.1","media":[],"coverageCount":0},{"id":"rel_6p84g-QSWHFHAwaGfKZFZ","version":"v0.23.0","type":"feature","title":"v0.23.0","summary":"## Claude Desktop\r\nClaude Desktop is now supported with Ollama Launch. \r\n\r\nClaude Cowork and Claude Code are supported within the Claude Desktop App.\r...","titleGenerated":null,"titleShort":null,"content":"## Claude Desktop\r\nClaude Desktop is now supported with Ollama Launch. \r\n\r\nClaude Cowork and Claude Code are supported within the Claude Desktop App.\r\n\r\n```\r\nollama launch claude-desktop\r\n```\r\n\r\n### Claude Cowork\r\n<img width=\"1272\" height=\"872\" alt=\"ca1\" src=\"https://github.com/user-attachments/assets/1d550e3f-0272-4429-8cb2-06d32344cb77\" />\r\n\r\n\r\n### Claude Code\r\n<img width=\"1272\" height=\"872\" alt=\"ca2\" src=\"https://github.com/user-attachments/assets/f2a5ed5f-3069-4975-bb22-ada82914a01c\" />\r\n\r\n\r\nClaude Code on the terminal can still be accessed through the CLI with:\r\n\r\n```\r\nollama launch claude\r\n```\r\n\r\n### Not supported yet\r\n- Web Search (coming soon)\r\n- Extensions \r\n\r\n## What's Changed\r\n* Launch Claude Desktop with `ollama launch claude-desktop`\r\n* The Ollama app now surfaces featured models from server-driven recommendations\r\n* Fixed OpenClaw gateway timeout on Windows by enforcing IPv4 loopback (thanks @UniquePratham)\r\n* Hardened Metal initialization to gracefully handle ggml kernel compilation failures\r\n\r\n## New Contributors\r\n* @UniquePratham made their first contribution in https://github.com/ollama/ollama/pull/15726\r\n\r\n**Full Changelog**: https://github.com/ollama/ollama/compare/v0.22.1...v0.23.0","publishedAt":"2026-05-03T03:34:11.000Z","fetchedAt":"2026-05-26T11:36:05.999Z","url":"https://github.com/ollama/ollama/releases/tag/v0.23.0","media":[{"type":"image","url":"https://github.com/user-attachments/assets/1d550e3f-0272-4429-8cb2-06d32344cb77","r2Key":"releases/5cfe3f59322ab80d1285ab69d2b0115a824a98df9326f525a30b4a693726749e.png","r2Url":"https://media.releases.sh/releases/5cfe3f59322ab80d1285ab69d2b0115a824a98df9326f525a30b4a693726749e.png"},{"type":"image","url":"https://github.com/user-attachments/assets/f2a5ed5f-3069-4975-bb22-ada82914a01c","r2Key":"releases/a13ddd0d01aee2fddcdcefc112709018c39dfe7f17ece89fe911b00c886b1b73.png","r2Url":"https://media.releases.sh/releases/a13ddd0d01aee2fddcdcefc112709018c39dfe7f17ece89fe911b00c886b1b73.png"}],"coverageCount":0},{"id":"rel_1ZKtOPVvNWaykFUOsTHdz","version":"v0.22.1","type":"feature","title":"v0.22.1","summary":"## What's Changed\r\n* Updated the **Gemma 4** renderer for thinking and tool calling improvements\r\n* Model recommendations are now updated without upda...","titleGenerated":null,"titleShort":null,"content":"## What's Changed\r\n* Updated the **Gemma 4** renderer for thinking and tool calling improvements\r\n* Model recommendations are now updated without updating Ollama\r\n* Aligned the desktop app's launch page with `ollama launch` integrations\r\n* Fixed the Poolside integration title in `ollama launch`\r\n\r\n\r\n**Full Changelog**: https://github.com/ollama/ollama/compare/v0.22.0...v0.22.1","publishedAt":"2026-04-28T20:30:57.000Z","fetchedAt":"2026-05-26T11:36:05.999Z","url":"https://github.com/ollama/ollama/releases/tag/v0.22.1","media":[],"coverageCount":0},{"id":"rel_hRiKeA8ttufHr1rI7VuPw","version":"v0.22.0","type":"feature","title":"v0.22.0","summary":"## New models \r\n* NVIDIA's [Nemotron 3 Omni](https://ollama.com/library/nemotron3)\r\n* Poolside's first open-weight coding model - [Laguna XS.2](https:...","titleGenerated":null,"titleShort":null,"content":"## New models \r\n* NVIDIA's [Nemotron 3 Omni](https://ollama.com/library/nemotron3)\r\n* Poolside's first open-weight coding model - [Laguna XS.2](https://ollama.com/library/laguna-xs.2)\r\n\r\n\r\n**Full Changelog**: https://github.com/ollama/ollama/compare/v0.21.2...v0.22.0","publishedAt":"2026-04-28T15:00:25.000Z","fetchedAt":"2026-05-26T11:36:05.999Z","url":"https://github.com/ollama/ollama/releases/tag/v0.22.0","media":[],"coverageCount":0},{"id":"rel_0FuemHP2Hn4-_LB7VU4ld","version":"v0.21.3-rc0","type":"feature","title":"v0.21.3","summary":"## What's Changed\r\n* api: accept \"max\" as a think value by @ParthSareen in https://github.com/ollama/ollama/pull/15787\r\n* openai: map responses reason...","titleGenerated":null,"titleShort":null,"content":"## What's Changed\r\n* api: accept \"max\" as a think value by @ParthSareen in https://github.com/ollama/ollama/pull/15787\r\n* openai: map responses reasoning effort to think by @ParthSareen in https://github.com/ollama/ollama/pull/15789\r\n\r\n\r\n**Full Changelog**: https://github.com/ollama/ollama/compare/v0.21.2...v0.21.3-rc0","publishedAt":"2026-04-24T12:15:38.000Z","fetchedAt":"2026-05-26T11:36:05.999Z","url":"https://github.com/ollama/ollama/releases/tag/v0.21.3-rc0","media":[],"coverageCount":0},{"id":"rel_okr9Ciq8ZxgmMf7Al0N5n","version":"v0.21.2","type":"feature","title":"v0.21.2","summary":"## What's Changed\r\n  * Improved reliability of the OpenClaw onboarding flow in `ollama launch`\r\n  * Recommended models in `ollama launch` now appear i...","titleGenerated":null,"titleShort":null,"content":"## What's Changed\r\n  * Improved reliability of the OpenClaw onboarding flow in `ollama launch`\r\n  * Recommended models in `ollama launch` now appear in a fixed, canonical order\r\n  * OpenClaw integration now bundles Ollama's web search plugin in OpenClaw\r\n\r\n## New Contributors\r\n* @madflow made their first contribution in https://github.com/ollama/ollama/pull/15733\r\n\r\n**Full Changelog:** https://github.com/ollama/ollama/compare/v0.21.1...v0.21.2\r\n","publishedAt":"2026-04-23T02:29:24.000Z","fetchedAt":"2026-05-26T11:36:06.033Z","url":"https://github.com/ollama/ollama/releases/tag/v0.21.2","media":[],"coverageCount":0},{"id":"rel_WLVfXGA6hF2zFLL1--9gR","version":"v0.21.1","type":"feature","title":"v0.21.1","summary":"## What's Changed\r\n### Kimi CLI\r\nYou can now install and run the Kimi CLI through Ollama.\r\n\r\n```\r\nollama launch kimi --model kimi-k2.6:cloud\r\n```\r\nKim...","titleGenerated":null,"titleShort":null,"content":"## What's Changed\r\n### Kimi CLI\r\nYou can now install and run the Kimi CLI through Ollama.\r\n\r\n```\r\nollama launch kimi --model kimi-k2.6:cloud\r\n```\r\nKimi CLI with Kimi K2.6 excels at long horizon agentic execution tasks through a multi-agent system.\r\n\r\n  * **MLX runner adds logprobs support** for compatible models\r\n  * **Faster MLX sampling** with fused top-P and top-K in a single sort pass, plus repeat penalties applied in the sampler\r\n  * **Improved MLX prompt tokenization** by moving tokenization into request handler goroutines\r\n  * **Better MLX thread safety** for array management\r\n  * **GLM4 MoE Lite performance improvement** with a fused sigmoid router head\r\n  * **Fixed model picker showing stale model** after switching chats in the macOS app\r\n  * **Fixed structured outputs for Gemma 4** when `think=false`\r\n\r\n**Full Changelog**: https://github.com/ollama/ollama/compare/v0.21.0...v0.21.1","publishedAt":"2026-04-22T00:18:02.000Z","fetchedAt":"2026-05-26T11:36:06.033Z","url":"https://github.com/ollama/ollama/releases/tag/v0.21.1","media":[],"coverageCount":0},{"id":"rel_cWBif3rKec7sF5yDdPAsh","version":"v0.21.0","type":"feature","title":"v0.21.0","summary":"## Hermes Agent\r\n\r\n```\r\nollama launch hermes\r\n```\r\n\r\nHermes learns with you, automatically creating skills to better serve your workflows. Great for r...","titleGenerated":null,"titleShort":null,"content":"## Hermes Agent\r\n\r\n```\r\nollama launch hermes\r\n```\r\n\r\nHermes learns with you, automatically creating skills to better serve your workflows. Great for research and engineering tasks.\r\n\r\n<img width=\"1329\" height=\"946\" alt=\"image\" src=\"https://github.com/user-attachments/assets/771d3383-95ed-4652-81e5-cf89514d25cc\" />\r\n\r\n## What's Changed\r\n\r\n- **Gemma 4 on MLX.** Added support for running Gemma 4 via MLX on Apple Silicon, including a text-only MLX runtime for the model. The MLX backend also picked up mixed-precision quantization, better capability detection, and a batch of new op wrappers (Conv2d, Pad, activations, trig, masked SDPA, and RoPE-with-freqs).\r\n- **Hermes and GitHub Copilot CLI in `ollama launch`.** Added both integrations, which can now be configured in one command alongside the rest of the supported coding agents.\r\n- **OpenCode moved to inline config.** `ollama launch opencode` now writes its config inline rather than to a separate file, matching how other integrations are handled.\r\n- **`ollama launch` no longer rewrites config when nothing changed.** Pressing → on a configured multi-model integration, or passing `--model` with the current primary, used to trigger a confirmation prompt and rewrite both the editor's config file and `config.json`. Now it's a no-op when the resolved model list matches what's already saved.\r\n- **Fixed `ollama launch openclaw --yes`** so it correctly skips the channels configuration step, so non-interactive setups complete cleanly.\r\n- **Restored the Gemma 4 nothink renderer** with the e2b-style prompt.\r\n- **Fixed the Gemma 4 compiler error** that was breaking Metal builds.\r\n- **Fixed macOS cross-compiles** so they no longer trigger `generate`, which was breaking cmake builds on some Xcode versions.\r\n- **Quieted cgo builds** by suppressing deprecated warnings during `go build`.\r\n\r\n**Full Changelog**: https://github.com/ollama/ollama/compare/v0.20.7...v0.21.0","publishedAt":"2026-04-16T22:00:17.000Z","fetchedAt":"2026-05-26T11:36:06.033Z","url":"https://github.com/ollama/ollama/releases/tag/v0.21.0","media":[{"type":"image","url":"https://github.com/user-attachments/assets/771d3383-95ed-4652-81e5-cf89514d25cc","r2Key":"releases/33f0685898d37bd68de59434f2346b3104a4e0f78be79d112da200718999af5a.png","r2Url":"https://media.releases.sh/releases/33f0685898d37bd68de59434f2346b3104a4e0f78be79d112da200718999af5a.png"}],"coverageCount":0},{"id":"rel_7jz2-9dpDDSScsdHczqnl","version":"v0.20.7","type":"feature","title":"v0.20.7","summary":"## What's Changed\r\n* Fix quality of gemma:e2b and gemma:e4b when thinking is disabled \r\n* ROCm: Update to ROCm 7.2.1 on Linux by @saman-amd in https:/...","titleGenerated":null,"titleShort":null,"content":"## What's Changed\r\n* Fix quality of gemma:e2b and gemma:e4b when thinking is disabled \r\n* ROCm: Update to ROCm 7.2.1 on Linux by @saman-amd in https://github.com/ollama/ollama/pull/15483\r\n\r\n**Full Changelog**: https://github.com/ollama/ollama/compare/v0.20.6...v0.20.7","publishedAt":"2026-04-13T23:33:00.000Z","fetchedAt":"2026-05-26T11:36:06.033Z","url":"https://github.com/ollama/ollama/releases/tag/v0.20.7","media":[],"coverageCount":0},{"id":"rel_FX_zEQrqliev6rldOXNQ4","version":"v0.20.6","type":"feature","title":"v0.20.6","summary":"## What's Changed\r\n* Gemma 4 tool calling ability is improved and updated to use Google's latest post-launch fixes\r\n* Parallel tool calling improved f...","titleGenerated":null,"titleShort":null,"content":"## What's Changed\r\n* Gemma 4 tool calling ability is improved and updated to use Google's latest post-launch fixes\r\n* Parallel tool calling improved for streaming responses \r\n* [Hermes agent](https://docs.ollama.com/integrations/hermes) Ollama integration guide is now available\r\n* Ollama app is updated to fix image attachment errors \r\n\r\n## New Contributors\r\n\r\n@matteocelani made their first contribution in [#15272](https://github.com/ollama/ollama/pull/15272)\r\n\r\n**Full Changelog**: https://github.com/ollama/ollama/compare/v0.20.5...v0.20.6","publishedAt":"2026-04-12T22:59:47.000Z","fetchedAt":"2026-05-26T11:36:06.033Z","url":"https://github.com/ollama/ollama/releases/tag/v0.20.6","media":[],"coverageCount":0}],"pagination":{"nextCursor":"2026-04-12T22:59:47.000Z|2026-05-26T11:36:06.033Z|rel_FX_zEQrqliev6rldOXNQ4","limit":20},"summaries":{"rolling":null,"monthly":[]}}