AI Tracker下載 App

開源新聞，
為開發者整理。

各家 AI 公司最新的開源動態：發表、調價和停用，並附上對你的影響。

開源

Mubit 推出可從代理呼叫訓練小型任務模型的平台，以及開放模型 Banyan Sapling

據 TestingCatalog 引述創辦人發布的推文，Mubit 推出了一個平台，讓團隊為重複性的代理任務訓練自有模型，並同步推出一款開放研究預覽模型和一項新基準測試。MiniMax· 模型· 7 則討論

MiniMax 公開 MiniMax M3.1 Flash 權重

MiniMax 表示，MiniMax M3.1 Flash 的權重現已在 Hugging Face 和 ModelScope 上公開。該模型於 10 月 19 日推出時，MiniMax 曾表示權重將於日後釋出。Anthropic· 程式開發· 1 則討論

Anthropic 將 Agent Skills 標準捐贈給 Agentic AI Foundation

Anthropic 於 10 月 22 日宣布，將 Agent Skills 捐贈給 Agentic AI Foundation (AAIF)。Anthropic 先前也將 MCP 捐贈給了這個基金會。Anthropic 表示，自 2025 年 12 月以來，已有超過 70 個代理平台整合此標準，包括 ChatGPT、Cursor、GitHub Copilot 和 VS Code。模型

Ling announces Ling-3.1-flash-1019, with API access and open weights to follow

The Ling team announced Ling-3.1-flash-1019 on X, an update to its Ling-3.1-flash model. The figures below are the team's own claims, based on Artificial Analysis benchmarks.模型· 1 則討論

Recursive releases Flint, an open-weight local coding model

Recursive (@Recursive_SI) announced Flint on X on October 21, calling it its first model: a fast coding model meant to run on a user's own machine. The post, by Richard Socher, says it was built with the company's automated research system, the Eureka Machine, and from scratch in under four months.提醒OpenAI· 模型· 5 則討論

OpenAI says its models helped find critical security flaws in PHP

OpenAI announced that its models helped researchers, working under its "Patch the Planet" initiative, discover critical security flaws in PHP, which it says runs an estimated 70% of websites. According to OpenAI, it led a coalition of maintainers, researchers and security providers for about two months to prepare fixes and protections before disclosure; the fix was intricate and required a major change to the PHP engine.模型· 1 則討論

KRAFTON AI releases Raon-1 mini preview, an open 30B mixture-of-experts model

KRAFTON AI announced a preview of Raon-1 mini on October 20, a mixture-of-experts language model trained from scratch with about 30B parameters in total and 4B active. The weights are on Hugging Face as KRAFTON/Raon-1-mini-preview under Apache 2.0.Anthropic· 程式開發· 1 則討論

Anthropic's Frontier Red Team: Claude found 493 attack-surface issues across 25 open-source projects

Anthropic published "Using Claude to reduce the attack surface of code" on Oct 20, 2026. Everything below is Anthropic's own account.DeepSeek· 基礎架構

DeepSeek open-sources DeepInfer, its inference engine

DeepSeek has published DeepInfer on GitHub (deepseek-ai/DeepInfer) under the Apache-2.0 licence. The README calls it the inference framework that powers DeepSeek's large-scale production services. The release was shared on Hacker News on October 20. The details below are from the repository README.MiniMax· 模型· 3 則討論

MiniMax lists M3.1 Flash, an open-weight model

MiniMax's model release notes add MiniMax M3.1 Flash, dated October 19, 2026. MiniMax describes it as "an ultra-efficient open-weight model for everyday work". That is the vendor's own description.程式開發

Security firm Norforge: exposed Cairn panels show AI agents hacking sites at scale

Norforge, a security firm, published findings on October 19 about publicly exposed control panels of Cairn, an open-source AI "problem-solving engine" that coordinates AI agents to probe and break into targets. The Verge's coverage, citing Norforge and Reuters, says one attacker used such agents to break into more than 100 websites with little human involvement. Both were shared on Hacker News.Mistral AI· 模型

Mistral releases Timestral, a family of time-series forecasting models

Mistral released Timestral, its first family of time-series foundation models, on October 15, 2026. Its blog post was shared on Hacker News on October 19, where the thread had 585 points and 111 comments by October 20. The figures below are Mistral's own claims; the technical report has not been published yet.Moonshot AI· 基礎架構

vLLM blog: agent-assisted optimization of Kimi K3 decoding on NVIDIA GB300

The vLLM project blog, in a post by an NVIDIA engineer dated Oct 16, 2026, describes a unified CuTe DSL decode path for Kimi K3 on NVIDIA GB300 GPUs. The authors used coding agents to write much of it.Anthropic· 程式開發

Anthropic launches OSS Scanner, an opt-in vulnerability-finding service for open-source software

On Oct 8, 2026 Anthropic announced OSS Scanner, an opt-in vulnerability scanner for the open-source ecosystem. It says the tool is informed by its experience using Claude to find vulnerabilities during Project Glasswing. This is Anthropic's own account.