What’s New in Mellea 0.4.0 + Granite Libraries Release
Details in article
Hugging Face
Build a Domain-Specific Embedding Model in Under a Day
Details in article
Hugging Face
Hugging Face Introduces SPEED-Bench for Speculative Decoding
Hugging Face introduced SPEED-Bench, a new unified and diverse benchmark specifically designed for evaluating speculative decoding techniques. Details...
Hugging Face
OpenAI Monitors Internal Coding Agents for Misalignment
OpenAI is implementing chain-of-thought monitoring to detect and mitigate misalignment risks in its internal coding agents. This approach involves ana...
OpenAI
Running Qwen 397B Locally on MacBook Pro with Apple’s ‚LLM in a Flash‘
Research demonstrates the ability to run a custom version of Qwen3.5-397B-A17B locally on a MacBook Pro M3 Max, achieving over 5.5 tokens/second. This...
Simon Willison Blog
Nvidia’s DLSS 5 is like motion smoothing for video games, but worse
Nvidia has unveiled DLSS 5, its latest upscaling technology, which the company touted as a major breakthrough in computer graphics. However, initial r...
The Verge AI
Open SWE: An Open-Source Framework for Internal Coding Agents
LangChain has launched Open SWE, an open-source framework tailored for building internal coding agents. Constructed upon Deep Agents and LangGraph, Op...
LangChain
Holotron-12B – High Throughput Computer Use Agent
Details in article
Hugging Face
Introducing LangSmith Sandboxes: Secure Code Execution for Agents
LangChain has unveiled LangSmith Sandboxes, providing secure code execution environments specifically designed for AI agents. Developers can now initi...
LangChain
Nemotron 3 Nano 4B: A Compact Hybrid Model for Efficient Local AI
Details in article
Hugging Face
State of Open Source on Hugging Face: Spring 2026
Details in article
Hugging Face
The First Healthcare Robotics Dataset and Foundational Physical AI Models for Healthcare Robotics
Details in article.
Hugging Face