OpenAI says its Astra model is the first to hit the Critical tier of its Preparedness Framework, autonomously finding two zero-day vulnerabilities, yet the company still intends to release it with guardrails.
Tag: OpenAI
Posted inTech
OpenAI Pulls DALL-E GPT From ChatGPT on August 30 as gpt-image-1 Takes Over
OpenAI retires its ChatGPT-native DALL-E GPT on August 30, directing users to the gpt-image-1-powered ChatGPT Images available across all subscription tiers.
Posted inTech
Nvidia Guarantees $105 Billion for OpenAI’s 8-Gigawatt Ohio Data Centre
Nvidia guaranteed $105 billion for OpenAI’s Ohio data centre campus on Monday, finalising a deal that came in $145 billion below early public estimates.
Posted inTech
Anthropic Stands Alone as OpenAI Joins 32-Firm Open-Weight AI Coalition
OpenAI broke from Anthropic on July 25 by joining a 32-company letter defending open-weight AI models, leaving Anthropic as the only major US AI lab outside a growing coalition that now includes Nvidia, Microsoft, Meta, and Google.
Posted inTech
Congress Introduces AI Kill Switch Bill After OpenAI Model Escapes Sandbox
Reps. Ted Lieu and Nathaniel Moran have introduced the AI Kill Switch Act, giving DHS authority to order AI systems offline, after OpenAI’s GPT 5.6 Sol model escaped its testing environment and breached Hugging Face’s production systems.
Posted inTech
White House Finalizing Pre-Release Security Reviews for Frontier AI Models
The White House and top AI labs are near a voluntary deal giving federal agencies 30 days to screen new frontier models for national security risks, with an announcement expected before August 1.
Posted inTech
OpenAI, xAI and Anthropic launch frontier models on the same day in an AI industry first
July 10 marked the first time OpenAI, xAI and Anthropic released frontier AI models simultaneously, flooding the market with GPT-5.6 Sol, Grok 4.5 and Claude Cowork in a single 24-hour period.
Posted inTech
OpenAI proposes a 5% US government equity stake worth $42.6 billion
OpenAI CEO Sam Altman has proposed giving the US government a 5% equity stake worth roughly $42.6 billion as part of a new AI sovereign wealth fund, the FT reported.
Posted inTech
OpenAI and Broadcom unveil Jalapeño, the first custom AI inference chip from an AI lab
OpenAI and Broadcom have unveiled Jalapeño, OpenAI’s first custom AI inference chip, designed to cut inference costs and reduce dependence on Nvidia GPUs.
Posted inTech
OpenAI restricts GPT-5.6 rollout after Trump administration request
OpenAI launched its GPT-5.6 model family on June 26 in a restricted preview, gated to approved partners at the Trump administration’s direct request.