Ollama v0.23.1 — Release Notes & Engineering Impact
Ollama v0.23.1. Source highlight: Gemma 4 MTP (Multi-token Processing) for the MLX runner
AI Models & LLM Frameworks archive — Page 5
1172 tracked entries. Older history is imported progressively; see source coverage.
Hugging Face Transformers: Release 5.8.0
Hugging Face Transformers v5.8.0. Source highlight: DeepSeek-V4 is the next-generation MoE (Mixture of Experts) language model from DeepSeek that introduces several architectural innovations over DeepSeek-V3. The architecture replaces Multi-head Latent Attention (MLA) with a hybrid local + long-range attention design, swaps residual connections for M
vLLM v0.20.1 — Release Notes & Engineering Impact
vLLM v0.20.1. Source highlight: This is a patch release on top of `v0.20.0` primarily focused on **DeepSeek V4 stabilization and performance improvements**, along with several important bug fixes.
Ollama v0.23.0 — Release Notes & Engineering Impact
Ollama v0.23.0. Source highlight: Claude Desktop is now supported with Ollama Launch.
LangChain langchain==1.3.0a1 — Release Notes & Engineering Impact
LangChain langchain==1.3.0a1. Source highlight: release(langchain): 1.3.0a1 (#37140)
LangChain langchain==1.2.17 — Release Notes & Engineering Impact
LangChain langchain==1.2.17. Source highlight: Changes since langchain==1.2.16
LangChain langchain==1.2.16 — Release Notes & Engineering Impact
LangChain langchain==1.2.16. Source highlight: Changes since langchain==1.2.15
Dify: v1.14.0
Dify 1.14.0. Source highlight: 🚀 What's New in v1.14.0?
Ollama v0.22.1 — Release Notes & Engineering Impact
Ollama v0.22.1. Source highlight: Updated the **Gemma 4** renderer for thinking and tool calling improvements
Hugging Face Transformers: Release v5.7.0
Hugging Face Transformers v5.7.0. Source highlight: Laguna is Poolside's mixture-of-experts language model family that extends standard SwiGLU MoE transformers with two key innovations. It features per-layer head counts allowing different decoder layers to have different query-head counts while sharing the same KV cache shape, and implements a sigmoi
Ollama v0.22.0 — Release Notes & Engineering Impact
Ollama v0.22.0. Source highlight: NVIDIA's Nemotron 3 Omni
vLLM v0.20.0 — Release Notes & Engineering Impact
vLLM v0.20.0. Source highlight: This release features 752 commits from 320 contributors (123 new)!
Open WebUI v0.9.2 — Release Notes & Engineering Impact
Open WebUI v0.9.2. Source highlight: 🧠 **PaddleOCR-vl document extraction.** Administrators can now use PaddleOCR-vl as a content extraction engine for document processing, with configurable API URL and token settings in document retrieval configuration. #23945
Hugging Face Transformers: Patch release v5.6.2
Hugging Face Transformers v5.6.2. Source highlight: Patch release v5.6.2
Hugging Face Transformers: Patch release v5.6.1
Hugging Face Transformers v5.6.1. Source highlight: Patch release v5.6.1
Ollama v0.21.2 — Release Notes & Engineering Impact
Ollama v0.21.2. Source highlight: Improved reliability of the OpenClaw onboarding flow in `ollama launch`
Hugging Face Transformers: Release v5.6.0
Hugging Face Transformers v5.6.0. Source highlight: OpenAI Privacy Filter
Ollama v0.21.1 — Release Notes & Engineering Impact
Ollama v0.21.1. Source highlight: You can now install and run the Kimi CLI through Ollama.
Open WebUI v0.9.1 — Release Notes & Engineering Impact
Open WebUI v0.9.1. Source highlight: 🐛 **Missing `aiosqlite` dependency.** Fixed a startup crash (`ModuleNotFoundError: No module named 'aiosqlite'`) when installing Open WebUI via `pip` or `uv` by adding the missing `aiosqlite` package to `pyproject.toml`. The dependency was listed in `requirements.txt` but not in the published packag
Open WebUI v0.9.0 — Release Notes & Engineering Impact
Open WebUI v0.9.0. Source highlight: ⚠️ **Database Migrations**: This release includes database schema changes; we strongly recommend backing up your database and all associated data before upgrading in production environments. If you are running a multi-worker, multi-server, or load-balanced deployment, all instances must be updated s
vLLM v0.19.1 — Release Notes & Engineering Impact
vLLM v0.19.1. Source highlight: This is a patch release on top of `v0.19.0` with Transformers v5.5.3 upgrade and bug fixes for Gemma4:
Ollama v0.21.0 — Release Notes & Engineering Impact
Ollama v0.21.0. Source highlight: ollama launch hermes
Ollama v0.20.7 — Release Notes & Engineering Impact
Ollama v0.20.7. Source highlight: Fix quality of gemma:e2b and gemma:e4b when thinking is disabled
Hugging Face Transformers: Patch release v5.5.4
Hugging Face Transformers v5.5.4. Source highlight: Patch release v5.5.4
Ollama v0.20.6 — Release Notes & Engineering Impact
Ollama v0.20.6. Source highlight: Gemma 4 tool calling ability is improved and updated to use Google's latest post-launch fixes
Ollama v0.20.5 — Release Notes & Engineering Impact
Ollama v0.20.5. Source highlight: OpenClaw channel setup with `ollama launch`
Hugging Face Transformers: Patch release: v5.5.3
Hugging Face Transformers v5.5.3. Source highlight: Small patch release to fix `device_map` support for Gemma4! It contains the following commit:
Hugging Face Transformers: Patch release: v5.5.2
Hugging Face Transformers v5.5.2. Source highlight: Small patch dedicated to optimizing gemma4, fixing inference with `use_cache=False` due to k/v states sharing between layers, as well as conversion mappings for some models that would inconsistently serialize their weight names. It contains the following PRs:
Hugging Face Transformers: Patch release v5.5.1
Hugging Face Transformers v5.5.1. Source highlight: Patch release v5.5.1
Ollama v0.20.4 — Release Notes & Engineering Impact
Ollama v0.20.4. Source highlight: mlx: Improve M5 performance with NAX
Ollama v0.20.3 — Release Notes & Engineering Impact
Ollama v0.20.3. Source highlight: Gemma 4 Tool Calling improvements
LocalAI v4.1.3 — Release Notes & Engineering Impact
LocalAI v4.1.3. Source highlight: fix(token): login via legacy api keys
LocalAI v4.1.2 — Release Notes & Engineering Impact
LocalAI v4.1.2. Source highlight: fix(autoparser): correctly pass by logprobs
LocalAI v4.1.1 — Release Notes & Engineering Impact
LocalAI v4.1.1. Source highlight: This is a patch release to address few regressions from the last release and the upcoming Gemma4, most importantly:
Ollama v0.20.2 — Release Notes & Engineering Impact
Ollama v0.20.2. Source highlight: app: default app home view to new chat instead of launch
LangChain langchain==1.2.15 — Release Notes & Engineering Impact
LangChain langchain==1.2.15. Source highlight: Changes since langchain==1.2.14
vLLM v0.19.0 — Release Notes & Engineering Impact
vLLM v0.19.0. Source highlight: This release features 448 commits from 197 contributors (54 new)!
LocalAI v4.1.0 — Release Notes & Engineering Impact
LocalAI v4.1.0. Source highlight: 🎉 LocalAI 4.1.0 Release! 🚀
Ollama v0.20.0 — Release Notes & Engineering Impact
Ollama v0.20.0. Source highlight: **Effective 2B (E2B)**
Hugging Face Transformers: Release v5.5.0
Hugging Face Transformers v5.5.0. Source highlight: Gemma 4 is a multimodal model with pretrained and instruction-tuned variants, available in 1B, 13B, and 27B parameters. The architecture is mostly the same as the previous Gemma versions. The key differences are a vision processor that can output images of fixed token budget and a spatial 2D RoPE to
LangChain langchain==1.2.14 — Release Notes & Engineering Impact
LangChain langchain==1.2.14. Source highlight: Changes since langchain==1.2.13
vLLM v0.18.1 — Release Notes & Engineering Impact
vLLM v0.18.1. Source highlight: This is a patch release on top of v0.18.0 to address a few issues:
Ollama v0.19.0 — Release Notes & Engineering Impact
Ollama v0.19.0. Source highlight: Ollama is now powered by MLX on Apple Silicon in preview
Dify: v1.13.3
Dify 1.13.3. Source highlight: 🚀 What's New in v1.13.3?
Hugging Face Transformers: Release v5.4.0: PaddlePaddle models 🙌, Mistral 4, PI0, VidEoMT, UVDoc, SLANeXt, Jina Embeddings v3
Hugging Face Transformers v5.4.0. Source highlight: Video Encoder-only Mask Transformer (VidEoMT) is a lightweight encoder-only model for online video segmentation built on a plain Vision Transformer (ViT). It eliminates the need for dedicated tracking modules by introducing a lightweight query propagation mechanism that carries information across fr
Open WebUI v0.8.12 — Release Notes & Engineering Impact
Open WebUI v0.8.12. Source highlight: 🌐 **Translation updates.** Translations for Simplified Chinese, Catalan, Portuguese (Brazil), Finnish, and Lithuanian were enhanced and expanded.
Open WebUI v0.8.11 — Release Notes & Engineering Impact
Open WebUI v0.8.11. Source highlight: 🔀 **Responses API streaming improvements.** The OpenAI proxy now properly handles tool call streaming and re-invocations in the Responses API, preventing duplicate tool calls and preserving output during model re-invocations. Commit, Commit, Commit, Commit
Ollama v0.18.3 — Release Notes & Engineering Impact
Ollama v0.18.3. Source highlight: Microsoft Visual Studio Code now directly integrates with Ollama via GitHub Copilot.
vLLM v0.18.0 — Release Notes & Engineering Impact
vLLM v0.18.0. Source highlight: Degraded accuracy when serving Qwen3.5 with FP8 KV cache on B200 (#37618)
LangChain langchain==1.2.13 — Release Notes & Engineering Impact
LangChain langchain==1.2.13. Source highlight: Changes since langchain==1.2.12