Every release that matters, side by side: new versions from the best open-source AI tools — models, runners, agents, image generators — and every new model that lands in the commercial APIs. Updated daily.
Loading…
Overview · llama.cpp v0.6.0 introduces the new llama_batch_ext extended batch API (with llama_process) for mixed token/embedding inputs and MTP/deepstack state
v0.31.0 · Highlights · This release features 717 commits from 307 contributors (96 new)! · * DeepSeek-V4.1-Flash performance: FlashMLA mega attention with the V
2.41.7 (2026-10-05) · Bug Fixes · * API: Let global members grant variable scopes to API keys (#40151) (cc25c71) · Features · * editor: Show Gateway credits pro
Change log · Improvements · * https://github.com/qdrant/qdrant/pull/10682 - Don't rewrite full sparse posting list on every upsert, making sparse vector upserts
Breaking Changes *none* · New Features *none* · Fixes * fix(backups): allow backups without collections by @tsmith023 in https://github.com/weaviate/weaviate/pu
The command-line tool for Vespa.ai. · Use it on Vespa instances running locally, remotely or in the cloud. Prefer web service API's to this in production. · See
What's Changed · Features * feat(mcp): support decision-model evaluators by @wochinge in https://github.com/langfuse/langfuse/pull/18166 * feat(experiments): an
1.824.0 (2026-10-05) · Features · * add an AI decision flow step (TypeSafe's Jev, Cloudflare's Clef) (#11474) (3cba8a7) * add notion oauth provider (#11528) (8b
Changelog · 🚀 Features core - 777ceef centralized HTTP retry mechanism in core (#19904), closes #19904 · worker - 0b35619 reach namespace files through the cont
Version 3.15.0 (October 5, 2026) · General · 1. 1aa68e1ca .mjz archives written on Windows now always use / separators. · 2. 1aa68e1ca Assets in a meshdir or te
Gymnasium v1.4.0 · This release adds max_concurrency to AsyncVectorEnv, support for arbitrary node and edge spaces in Graph, optional key sorting for Dict, gene
What's Changed * feat(web,local): add wordmark, Windows overlay scrollbars, and marketing site polish by @AnishSarkar22 in https://github.com/MODSetter/SurfSens
LoopX v1.2.4 · More reliable Chat coordination, explicit local state migration, and scoped preferences for long-running agents. · State Kernel & Control Plane ·
What's Changed ⚠️ Compatibility Notes * Handle draft-7 list-form items in JsonSchemaTransformer, TestModel and Mistral streamed output, and make TestModel accep
OpenClaw v2026.9.8 · 58 commits · 43 pull requests · 21 contributors · Changes included in this release · The release notes and changelog contain the same conte
Feature · * ocr: Allow configuring RapidOCR model size (#4500) (a1c5ff2) · Fix · * opendocument: Read the chart object's own title (#4463) (4c109c2) * code-form
Verify Docker Image Signature · All LiteLLM Docker images are signed with cosign. Every release is signed with the same key introduced in commit 0112e53. · Veri
What's Changed * fix(mcp): derive model output schemas from serialization by @asts-top in https://github.com/ag2ai/ag2/pull/3317 * fix(tenki): upgrade sdk and p
🎉 LocalAI 4.11.0 Release! 🚀 · · LocalAI 4.11.0 is out! · This release makes LocalAI more useful for audio understanding, resilient model serving, structured dec
Highlights · *779 PRs from 227 contributors.* · New models · | Model | Type | Cookbook | |---|---|---| | DeepSeek-V4.1 Flash | LLM / VLM | link | | GigaChat 3.5
Existing users get a one-time What's new dialog introducing Connectors, with connector logos, a clickable example prompt, and an Open Connectors button. You can
Paperclip v2026.1001.0 · > Released: 2026-10-01 · Paperclip v2026.1001.0 carries 77 commits, promoted from 2026.921.0-beta.1. The headline is reliability: a dee
Changelog · New Features · - Live progress & cancellation for page sync: Knowledge.stream_sync_pages() / astream_sync_pages() yield typed PageSyncProgress snaps
✨ Features · - Resizable Picture-in-Picture window for MCP Apps #12417 - Fetch model metadata from models.dev live with bundled fallback #12560 - Quoting action
2.11.0 (2026-10-01) · Highlights · This release adds graceful cancellation and tool confirmation to workflows, a tool for consulting another model mid-task, and
What's new · - Built with WXT: the extension is now a single package built with WXT instead of a multi-package monorepo. Development mode reloads pages instantl
Hexagon HTP · Microsoft-signed HTP catalog (llama.cpp @ 942561). No cert import required on Windows on Snapdragon. · What's Changed * docs(build): add instructi
Highlights · * Ray Data LLM & Ray Serve LLM are GA/Stable: the LLM APIs graduate to general availability this release (\#65194), alongside an upgrade to vLLM 0.
Release: v2.26.3 [skip ci] Patch · - Make: Tune only tests & benchmarks to the host, on every compiler (74576b3)
Highlights · * Ray Data LLM & Ray Serve LLM are GA/Stable: the LLM APIs graduate to general availability this release (\#65194), alongside an upgrade to vLLM 0.
What's Changed * fix(collector): read spreadsheet dates as dates, not serial numbers by @L4XB in https://github.com/Mintplex-Labs/anything-llm/pull/6478 * Drop
⭐️ Highlight · ⚡️ Faster SentenceWindowRetriever SentenceWindowRetriever now queries the Document Store once per run or run_async call instead of once per retri
This release brings faster navigation, shareable run settings, and clearer errors to Unsloth Desktop. It also 4x speeds up Laya decisions, expands hosted Decisi
What's Changed · Version bumped from 1.0.54 to 1.0.55 · Recent Changes - Bump version to 1.0.55 (#8758) (d45e60095) - SKY-17491: Inline run reliability below an
This is a PEFT release fixes an issue that prevented encoder-decoder models to work when using Transformers ≥ 5.18.0. · Changes: · - #3845
llmware 0.4.7 release wheel includes prepackaged, preconfigured llama cpp binaries for multiple optimized environments 'out of the box', including: DGX Spark, A
20.19.0 (2026-10-01) · Features · * datagen: send each recorded application to its own project (#16682) (d99a4db) * tracing: add the DECISION span kind across t
New · - Improved the performance of declarative automation for partitioned assets with a long materialization history. - [ui] The UI now has a mobile layout for
Axolotl v0.20.0 Release Notes · There has been 67 commits since v0.19.0 (September 10, 2026). · We have added a bunch of features including GGUF export, Ringmas
🔥 What's New? · New Features · - #14209 - AutoPilot action gating modes - Ask First, Auto, Unsupervised (by @Pwuts) - #14817 - Workflows started from AutoPilot
Open Interpreter 0.0.55 · Models · Fix offline model listing for configured non-OpenAI providers. When an endpoint has no matching catalog identity, the app ser
Core · Bugfixes - Send namespaced session and parent-session identity headers with model requests. - Re-sign locally compiled macOS binaries so they run reliabl
This week's release includes BYOK support for GPT-6.1 Sol, on-the-fly model selection for new subagents, automatic subagent compaction, a unified/split toggle f
🐞 Bug Fixes · * fix(engine): remove per-sandbox piece LRU (#15786) @amrdb · This restores the piece-loading fix from 0.91.3, which 0.92.0 was missing. · > 📌 Upg
This minor release focuses on improving the simulation speed on GPU for non-batched to moderate batch sizes (~4000) while reduce memory usage across the board.
This release is meant to fix the following regressions and silent correctness issues: · Silent correctness fixes - Fix incorrect torch.linalg.lstsq solutions on
Clef decision models · Ollama now supports Clef and Clef Flash, Cloudflare's new open-source decision models, through /v1/systemone. · Clef (27B) and Clef Flash
What's Changed * Improve qwen 2.1 KV cache location logic by @rattus128 in https://github.com/Comfy-Org/ComfyUI/pull/16429 * Compile Qwen Image 2.1 transformer
Changes since langchain-core==1.6.5 · release(core): 1.6.6 (#40906) fix(anthropic): support Claude Sonnet 5.5 compatibility (#40882) docs(core): fix docstring e
What's Changed 🐛 Bug Fixes * fix: reject mixed blocked DNS answers with IP allowlists by @erichare in https://github.com/langflow-ai/langflow/pull/15334 * fix(a
RAGFlow v1.0.0-rc1 is a preview release of v1.0.0. This release marks a comprehensive rewrite of RAGFlow in Go, delivering a substantial reduction in system res
What's Changed · * Fix the server-mode crash at the first weight sync on vLLM 0.20 to 0.25 by @albertvillanova in https://github.com/huggingface/trl/pull/7410 *
Minor Changes · - cdabe06: Allow the Google and Cloudflare providers' executeToolCall and the OpenAI Responses provider's handleResponse to execute through a su
更新内容 / What's Changed · 修复 / Fixes - WebUI：非安全上下文（如 HTTP 或非 localhost 访问）下 crypto.randomUUID 不可用时提供 UUID 回退实现，避免预览等功能失效 / WebUI: fall back to an alternative UUI
## ⚠️ Upgrade notes · - pip installs: rotate your SECRET_KEY if other users on the machine can read your data directory. Older versions saved the generated SECR
This patch release includes important bug fixes to ONNX build. · What's Changed * Validate offset in save_external_data function PR Raised https://github.com/on
Signed and notarized PuppyOne Desktop release. · Commit: 0c666447b0d7c8811de592c530a03f52fb030595 Promoted Internal release: v0.3.27-internal.48 R2 version pref
What's Changed 🚀 Features * [Feat]: Support output input logprobs by @RunningLeon in https://github.com/InternLM/lmdeploy/pull/4793 * Expand SM90 quantized GEMM
What's Changed · Features - Add support for native Gemini 3.8 Flash - Implement evaluation of the last traced run through AMP in crewai eval - Record the last t
Bug Fixes · - browser_find takes filename to save matches to a file instead of the response (microsoft/playwright#42077). (microsoft/playwright#42968) - browser
Added · - Export and import background jobs (dataset, annotations, backup) are now retried automatically on failure; the number of retries and the delay between
This release fixes a number of bugs in version 6.14.1 and adds a couple of useful new features. · New Features in this release * You can now import custom fonts
What's Changed * fix(html): extract definition lists instead of discarding them by @r0h1tb in https://github.com/Unstructured-IO/unstructured/pull/4502
koboldcpp-1.122.1 · https://github.com/user-attachments/assets/8f9448d8-8e35-4269-bae9-c3122bdfb8c3 · *PSA: KoboldCpp needs YOUR help to fight a fake phishing s
1.24.0 (2026-09-25) · What's Changed Features * feat: toggle all workspace folders from the Conversations header by @kuldeepsinh19 in https://github.com/OpenHan
DSPy 3.4.0 · DSPy 3.4 introduces Jev integration through TypeSafe, with three experimental decision types—Noul, Choice, and Score—that bring probability evidenc
Bug Fixes: - Config (OSS): ConfigManager no longer injects OpenAI's default baseURL and model into non-OpenAI LLM providers. The defaults now apply only to open
What's Changed * Update dependency huggingface-hub to ~=1.28.0 by @renovate[bot] in https://github.com/containers/ramalama/pull/2911 * Find short HTTP downloads
v1.23.5 - 2026.09.24 · Downloads · - macOS (Apple Silicon) - macOS (Intel) - Windows (x64) - Linux (x64 DEB) - Linux (ARM64 DEB) · What’s Changed · 1. Fixed ext
What's Changed * Change default package feed index. by @dworthen in https://github.com/microsoft/graphrag/pull/2483 * Fix/cache event loop lifecycle by @dworthe
Hermes Agent v0.21.5 (v2026.9.24) · Release Date: September 24, 2026 · > Patch release. This tag rolls up the ~460 PRs merged since v0.21.4 into a stable tagged
Changes since cli==0.4.32
Agent Zero v2.13 · Highlights · This release brings a significantly improved file browser, sidebar project folders, slash commands for messaging integrations, n
🎉 Crawl4AI v0.9.4 Released! · 📦 Installation · PyPI: · Docker: · Note: Docker images are being built and will be available shortly. Check the Docker Release wor
torchaudio is now an optional dependency instead of required. Only torch is required — the model itself never used torchaudio, only the read_audio / save_audio
Added · - 📉 Far smaller slim image. A slim build now comes down at around 175 MB, near enough 89% smaller than the last release, the local models, the packages
Release Notes · [2026-09-21] · llama-index-agent-agentmesh [0.3.0] · - fix: resolve a ton of security alerts (#22855) · llama-index-agent-azure [0.4.0] · - fix:
This release rolls up dozens of small patches and bug fixes. For typical inputs and use cases, we expect outputs and behavior to remain largely unchanged from v
Published from https://github.com/lobehub/lobehub/pull/19706 on merge commit 14dfc07b. · Full Changelog: https://github.com/lobehub/lobehub/compare/v2.2.17...v2
v3.0.2 · Release date: September 20, 2026 · | Milvus Version | Python SDK Version | Node.js SDK Version | Java SDK Version | Go SDK Version | | -------------- |