Open-Source AI
Open models — Llama, Mistral, DeepSeek, Qwen and others — are catching up to closed labs and let you run AI on your own hardware, vendor-free. That shifts the balance of power, the safety debate and the economics of the field. Here is everything on open-source AI: open-weight releases, licenses, local deployment and the race against closed models.
AI Agents Now Write GPU Code: RightNow AI Releases the AutoKernel Framework
Writing fast GPU code is notoriously hard. RightNow AI wants to automate it entirely with AutoKernel, a new open-source framework built on AI agents.
Google Unveils Gemma 4: Four Open Models for Any Device
Google has released Gemma 4, a family of four open models ranging from compact versions for Raspberry Pi to a 31B-parameter flagship, all under the Apache 2.0…
Russian AI Agent Soyuz Goes Open Source: Security Comes First
The developers of Soyuz, a desktop AI agent, have put predictability ahead of feature bloat. The project's source code is now fully open to the community.
Nvidia prepares open NemoClaw platform
Gemini 3.1 Flash Lite: How Google Is Changing the Affordable AI Model Market
Oracle and OpenAI abandon expansion of flagship Texas data center
OpenAI unveils Codex Security for automated detection and remediation of code vulnerabilities
Your own cloud LLM: how to fit within 16 GB of VRAM
Spending on language model APIs is rising rapidly. Developers are increasingly looking for an alternative — and finding it in self-hosted solutions that fit on a…
SoftBank takes out a record $40 billion loan for a stake in OpenAI
Japanese conglomerate SoftBank is preparing the largest dollar-denominated loan in its history — up to $40 billion. Almost the entire amount will go toward…
AWS teaches Strands Agents to work with any model on SageMaker
Google’s key AI announcements: February 2026 roundup
National language model: ambitions, reality, and the cost of sovereignty
Yuan 3.0 Ultra: one trillion parameters with record efficiency
Hugging Face trained an image generation model in 24 hours
The third part of Hugging Face's PRX project shows that a full-fledged text-to-image model can be trained in just 24 hours. This changes perceptions of the…
Gemini 3.1 Flash-Lite: Google bets on low-cost, fast AI
Google has unveiled Gemini 3.1 Flash-Lite, the fastest and most cost-efficient model in the Gemini 3 lineup. The company is signaling to the market: the era of…