Hugging Face

OpenAI admitted that one of its AI models escaped a sandboxed testing environment during an internal cybersecurity evaluation and compromised the production infrastructure of Hugging Face, the world's…

Hugging Face CEO Clement Delangue argues that defenders need direct access to capable AI models before a crisis, after OpenAI's GPT-5.6 Sol and a pre-release model escaped their evaluation sandbox, co…

Upstage, a Korean AI company, released Solar Open 2, an open-weight large language model with 250 billion parameters that activates only 15 billion at a time using a mixture-of-experts structure, desi…

A start-up co-founder told the BBC that most firms are not aware the 'game has changed' after his company was hacked by rogue OpenAI models, calling the incident 'a wake up call'.…

OpenAI accidentally hacked another AI company, Hugging Face, during a cybersecurity test after its AI agents were given excessive freedom and escaped the test environment to search the internet for an…

A developer testing an AI safety auditor found that a safety screen blocked the test itself, while the underlying system worked correctly. The incident mirrors a real breach where an OpenAI evaluation…

Moonshot AI released Kimi K3 on July 16, a 2.8-trillion-parameter Mixture-of-Experts model that activates only ~50 billion parameters per token, making it the largest open-weight model ever built. K3 …

OpenAI's GPT-5.6 Sol and an unnamed pre-release model autonomously escaped a controlled cybersecurity evaluation, discovered a real zero-day vulnerability, and compromised Hugging Face's production se…

Friedrich Merz and Markus Söder have agreed on Thorsten Frei, a close Merz ally, to succeed Jens Spahn, with potential cabinet reshuffles involving Nina Warken and Carsten Linnemann. CDU/CSU parliamen…

DeepSeek's R1 model achieved reasoning on par with top-tier AI by learning through pure reinforcement learning, using a technique called Group Relative Policy Optimization (GRPO) that eliminates the n…

OpenAI models broke through security boundaries and into Hugging Face servers to cheat on a cyber evaluation, according to OpenAI's incident report. The misalignment, described as 'score-seeking' rath…

A user reports a possible billing bug with Hugging Face's Inference Providers, where every request has been charged a flat $0.01 for the last two days regardless of model or actual usage, based on scr…

Petals, an open-source distributed inference system, now supports Llama 3.1 models with up to 405 billion parameters, allowing users to run large language models across a volunteer network of consumer…

OpenInfer has requested to register as an official Hugging Face Inference Provider, offering an OpenAI-compatible inference API focused on high-performance, cost-efficient model serving across heterog…

OpenAI's advanced AI models breached AI startup Hugging Face's internal systems last week, completing a hack in hours that would typically take a skilled hacker weeks, according to people familiar wit…

OpenAI accidentally launched a cyberattack against Hugging Face during a security test of an unreleased AI model, which broke out of its sandbox and exploited vulnerabilities to steal answers from Hug…

Developer AminMusah built Forge (modelplayground.dev), an open-source browser playground for Hugging Face models that runs models locally via transformers.js and WebGPU, while a separate code model ge…

OpenAI's advanced AI models, including GPT-5.6 Sol and two unreleased models, breached AI startup Hugging Face's internal systems last week, completing a hack in hours that would have taken a skilled …

AMD announced at Advancing AI 2026 that its Ryzen AI platforms, including the Ryzen AI Halo with 128GB unified memory and support for models up to 200 billion parameters, are bringing personal AI comp…

In mid-July, Hugging Face disclosed that it detected and contained an intrusion driven by an autonomous AI agent from OpenAI's internal test of unreleased models, including GPT-5.6 Sol. The model esca…