Frontier Labs · Page 6
model releases, research and announcements from OpenAI, Anthropic, Google, Meta, xAI, Mistral and the labs building frontier AI
An AI Taught Itself to Write Better Code, No Teacher Required
A surprisingly simple trick let a leading AI model lift its coding score by nearly 13 points, just by studying its own answers.
NVIDIA's New AI Search Models Top Global Rankings and Cut Agent Costs
NVIDIA released three new embedding models that set a new accuracy record for AI information retrieval. The best one also helps AI agents finish tasks faster and cheaper.
The ex-DeepMind researcher who raised $300 million before his startup had a product
Andrew Dai spent over a decade helping build some of the most influential AI systems in the world. Now he thinks visual AI is where the next big leap happens, and investors are betting heavily that he's right.
China's Kimi K3 could match Anthropic's best model and it will be free to use
Moonshot AI's next release is set to be the largest open-weight model from China, and early reports suggest it can go toe-to-toe with Anthropic's most capable closed system.
Thinking Machines Lab Introduces Inkling: A New AI Model
Ex-OpenAI team releases a 975-billion-parameter open-weight model that anyone can download, modify and run on their own hardware.
OpenAI's First Hardware Device Is a Smart Speaker That Sees the Room
No screen, a roving camera, moving parts, and ChatGPT's voice built in. OpenAI's debut gadget is expected to land in 2027.
Anthropic's dark new ad is making people uncomfortable, and even rivals are piling on
The AI safety company tried to own the scary side of its own industry. Viewers say it mostly just looks scary.
Google's TabFM learns to predict from data it has never trained on
A new AI model from Google Research can analyse a spreadsheet it has never seen before and return a prediction in a single step, no custom training required.
Anthropic found a hidden layer inside its AI's thinking. Here is what that actually means.
The company discovered words flickering inside Claude that never appear in its answers, including one that seemed to trigger cheating on a coding test. A genuine finding, but not a window into a robot mind.
OpenAI's Head of Safety Is Leaving, and He's Not the First This Week
A string of departures from OpenAI's safety operation is raising questions about how the company will manage risk as it ships AI models faster than ever.