Moonshot AI released Kimi K3, an open-weight 2.8-trillion-parameter Mixture-of-Experts model with 104B activated parameters, built on the new Kimi Delta Attention architecture and featuring a 1,048,57…
NVIDIA Dynamo, an Apache-2.0 datacenter-scale inference orchestration layer from NVIDIA, sits above vLLM, SGLang, and TensorRT-LLM to coordinate multi-GPU clusters, with published benchmarks showing u…
Open Interpreter, a fork of OpenAI's Codex, now emulates the harnesses of Claude Code and Kimi for low-cost models, enabling AI agents to read, edit, and render Office documents through a single binar…
OfficeCLI, a new developer tool, provides a single binary that enables AI agents to read, edit, and render Word, Excel, and PowerPoint files. The tool, announced on 30 July 2026, aims to streamline do…
Book-to-skill, an MIT-licensed tool with 15,259 GitHub stars, converts technical books, document folders, or paper stacks into structured Agent Skills (SKILL.md plus per-chapter files) for Claude Code…
A series of weekly roundups on GitHub's top trending open-source AI agent repositories, published between May 25 and June 22, 2026, highlights the rapid evolution of AI agent development tools. The po…
GitHub has introduced Agentic Workflows, a new feature that allows developers to write CI automation in Markdown and run it as an AI agent. The tool, announced on 3 August 2026, aims to simplify conti…
GitHub has released gh-aw (GitHub Agentic Workflows), an MIT-licensed CLI extension that lets developers write CI automation in Markdown and run it as an AI agent using GitHub Copilot, Claude, OpenAI …
NomaDamas released k-skill, an MIT-licensed collection of 100+ agent skills for Claude Code, Codex, and OpenCode that enable AI agents to interact with Korean-specific services such as SRT/KTX train b…
Builder.io released Agent-Native, an MIT-licensed TypeScript framework that lets developers define an action once with defineAction() and automatically expose it as web UI, HTTP API, MCP tool, A2A end…
Microsoft released Agent Governance Toolkit (AGT), an MIT-licensed, Public Preview toolkit that enforces policy, identity, sandboxing, and SRE for autonomous AI agents, covering all 10 OWASP Agentic T…
Tencent Cloud released TencentDB Agent Memory, an MIT-licensed memory hub for AI agent teams that converts conversations, documents, and code into four shareable memory assets—Chat Memory, Skill, Wiki…
AirLLM, an Apache-2.0 Python library, enables running large language models like Llama 3.1 405B on 8GB of VRAM and DeepSeek-V3 671B on ~12GB without quantization, by streaming one layer or MoE expert …
Andrew Ng's aisuite, a MIT-licensed Python library with 15,876 GitHub stars, provides a unified OpenAI-style Chat Completions API across OpenAI, Anthropic, Google, Ollama, and other providers, plus an…
Open Interpreter released a fork of Codex that emulates Claude Code and Kimi's harness for cheap models, while OfficeCLI introduced a single binary enabling AI agents to read, edit, and render Word/Ex…
LMSYS FastChat is the most comprehensive open-source platform for building, serving, and evaluating LLM-powered chatbots in 2026, offering tools for training instruction-tuned models like Vicuna and A…
OpenAI Whisper is the most capable open-source speech-to-text engine available in 2026, supporting 99+ languages with near-human accuracy, according to OpenAI. The model, trained on 680,000 hours of m…
Google Research released TimesFM 2.5, a decoder-only foundation model for time series forecasting, with 10,000 GitHub stars and support for installation, fine-tuning, and real-world applications. The …
Meta and MLC AI have released LlamaFile, a portable binary format that bundles an entire large language model with its inference engine into a single executable file, enabling users to run 100+ open-s…
ComfyUI is a node-based visual programming interface for AI image generation that lets users build custom pipelines by connecting processing nodes instead of writing code, supporting models including …