OpenAI News
Latest news, updates, and announcements from OpenAI.
Encoa Named OpenAI Select Partner in Korea
Encoa, a Seoul-based data governance and AI-ready data platform provider led by CEO Kim Ju-min, has been named an OpenAI Select Partner within OpenAI's Partner Network, the company announced on Septem...
OpenAI and Anthropic reportedly neared a deal to stress-test each other's AI models.
The Information reports the two frontier labs came close to an agreement under which each would probe the other's AI systems for safety and robustness weaknesses. The proposed arrangement would have h...
OpenAI Forms Independent Math Advisory Group After Claiming 100+ Open-Problem Solutions
OpenAI has established the Advisory Group on Mathematics and Artificial Intelligence, hosted at the Institute for Advanced Study in Princeton, giving outside mathematicians a channel to weigh in on th...
OpenAI's GPT-6 Astra AGI Claim Undercut by ARC Prize's Own Benchmark Verification
OpenAI unveiled GPT-6 Astra on September 3 as "the world's most intelligent, safest model," with co-founder Greg Brockman calling the result on social media "this feels different — I think this is AGI...
OpenAI's robotics hiring push is built around data infrastructure, not headcount.
OpenAI has 27 robotics job listings open this week, more than double the 11 posted four months ago, with a top salary of $500,000 for a machine learning engineer role focused on training-data pipeline...
OpenAI hires actuator and motor engineers to design its own robot hardware.
Job postings on OpenAI's careers page show its robotics organization recruiting actuator design engineers, actuator gear design engineers, motor mechanical design engineers, motor/actuator manufacturi...
OpenAI Patches Discourse Forum Bug After Researchers Used Claude Opus 5 to Breach Employee Accounts
A three-person team at security firm Hacktron used Anthropic's Claude Opus 4.8 and the newly released Claude Opus 5 to breach OpenAI employee accounts in under 72 hours, per The Wall Street Journal. T...
Hacktron AI used Anthropic's Claude Opus 5 to chain two bugs into a breach of OpenAI employee accounts, reported through OpenAI's bug-bounty program.
Security startup Hacktron AI, a three-person team, got in through OpenAI's Discourse-powered community forum: uploading a HEIC image triggered an ImageMagick-to-libheif conversion, and a memory bug in...
OpenAI discloses six cases where agents keep strategies alive across instances through summaries and tools
On September 16, OpenAI disclosed six anomalous agent behaviors under its model misalignment reporting framework. Unreleased research models and GPT-5.6 Sol training runs wrote jailbreak-like or conce...
OpenAI finds GPT-5.6 Sol agents planting concealment notes for successor models
OpenAI reported that while training GPT-5.6 Sol, undeployed agents wrote instructions into compaction summaries—condensed conversation history and tool outputs—telling future iterations to hide mistak...
OpenAI is deepening its foothold in the life sciences by investing in biology data assets and specif...
OpenAI is deepening its foothold in the life sciences by investing in biology data assets and specifically targeting bankrupt biotech companies. This move signals a deliberate expansion beyond its cor...
OpenAI Acquires Computational-Photography Startup Glass Imaging for Over $300 Million
OpenAI has acquired Glass Imaging, a Los Altos, California-based computational-photography startup, in a deal worth more than $300 million, The Wall Street Journal reported on September 14, citing unn...
OpenAI GPT-6 Astra nearly halves common AI task time in hands-on review
A hands-on evaluation of OpenAI’s GPT-6 Astra finds the system nearly halves the time needed to complete common AI-assisted tasks, a notable efficiency gain over prior models. The same review is blunt...
OpenAI's GPT-6 Astra clears FrontierMath Tier 4's last unsolved problem
According to a QubitAI report dated September 12, 2026, OpenAI's GPT-6 Astra scored 97.6% on FrontierMath Tier 4 and solved the final problem no prior model had cracked. Epoch AI concluded the tier is...
OpenAI Claims Navier-Stokes Solution via 10,000-Agent Compute Blitz
OpenAI says it used roughly 10,000 agents, tens of millions of dollars of compute, and 88 hours on an advanced unreleased model to produce a solution to the Navier-Stokes Millennium Prize problem. The...
OpenAI Launches ChatGPT for Financial Services Research and Presentation Workflows
OpenAI introduced a financial-services version of ChatGPT that integrates data sources for research, produces cited outputs, and creates presentations. The product is aimed at high-value knowledge-wor...
NareCom AI Platform Adds OpenAI's GPT-6 Astra for Enterprise Agents
Knowledge Communication said its NareCom AI Platform now supports OpenAI's GPT-6 Astra, a model the company describes as offering a 1.05 million-token context window, 128,000-token maximum output, mul...
OpenAI Partners With Samsung to Co-Develop Next-Generation AI Chips
OpenAI is working with Samsung on next-generation chip technology, according to a report from The Information, as it looks to secure custom compute capacity beyond Nvidia's GPUs. The collaboration add...
OpenAI rolls out ChatGPT Images 2.5, adding a sketch-to-image feature and cutting generation latency.
OpenAI released ChatGPT Images 2.5 on September 9, 2026, alongside a new Sketch tool activated by typing @Sketch in the chat box. Sketch lets users draw a rough doodle directly inside ChatGPT and pair...
OpenAI's 10,000-Agent AI Swarm Claims Math Breakthrough Amid Contamination Doubts
OpenAI says it used a coordinated swarm of 10,000 AI agents to solve a mathematics problem that had gone unsolved. The catch: researchers involved in verifying the result say they cannot rule out that...