OpenAI just teamed up with the Frontier Model Forum and member firms on a new framework to identify when bypassing AI safeguards triggers serious cyber harm. The goal is a shared standard to measure unlocked capabilities and ease of misuse, helping security teams prioritize where to build stronger defenses.
OpenAI teamed up with biotech firm Ligo to launch BioBasicsBench, evaluating how AI agents interpret raw data across 100 routine wet-lab tasks. Running at maximum reasoning effort, GPT-6 Astra scored 66.7%, comfortably beating GPT-5.6 Sol's 43.8%. Still, the models struggle with control checks and colony counting, showing AI lab assistants aren't ready to fly solo just yet.
OpenAI just announced it will stop using non-disclosure agreements with public officials for future data center projects, following voter pushback over rapid tech expansion. The company also pledged to cover the full cost of its electricity and related infrastructure, while publishing local job creation data. Amazon made a similar NDA pledge in October.
OpenAI just expanded access to its Daybreak security platform for organizations of all sizes, independent researchers, and open-source maintainers. The program includes Daybreak Blue for defensive validation with fewer refusals and Daybreak Red for authorized testing, alongside a broadened GPT-6 Astra alpha rollout.
OpenAI just launched GPT-6 Cyber, a specialized cybersecurity model designed to find, exploit, and patch software vulnerabilities without wandering out of scope. In a single automated run, it validated and patched 948 Linux kernel privilege escalation bugs at a $1.2 million API cost. It is currently available to select Daybreak Red users.
OpenAI just announced it will acquire Pierre Computer Company to upgrade Git workflows and code review inside Codex. Pierre's Trees and Diffs libraries already power file navigation and change reviews in the Codex app, and its open-source tools will remain open. Why it matters: as software teams deploy multiple coding agents in parallel, OpenAI wants to redefine modern version control for the agentic era.
OpenAI just launched Images 2.6 across ChatGPT, Codex, and its API, boosting generation speeds by 50% and adding native support for animated GIFs. Developers get two API variants—including the six-second Flare model—plus SVG exports and fine-grained scene complexity controls. It is live now for all subscription tiers.
OpenAI just revealed it's working with Chip Ganassi Racing engineers to turn 3D scans into editable CAD models and manufacturing drawings. The company is also collaborating with Mazda to explore new vehicle designs through simulation, relying on partner engineers to review outputs and sharpen the models.
MIT researchers found that OpenAI’s GPT-6 Astra can inspect and directly rewrite neural network weights without any gradient descent or retraining. In one test, Astra modified just 0.016% of a small Qwen model’s weights to eliminate repetitive loops, bumping its math score from 7.9% to 48.1%. If frontier models can reliably program weight space, AI self-improvement might arrive much faster than expected.
OpenAI just revealed how its frontier models are tackling physical engineering, with two staffers using GPT-6 Astra to design a hardware prototype in under a week. Internal hardware token usage has jumped 23-fold since February, and partners like Mazda and Chip Ganassi Racing are already testing CAD conversion and aerodynamics simulations. It's not fully autonomous yet, though—prototypes still required human review to fix manufacturing tolerances.
OpenAI just announced an upcoming internal security sprint alongside Cathedral, putting its own models to work to probe and harden company systems. It is also quadrupling bounties for critical bug findings and pledging $1 billion in subsidized access for frontline defenders over six months. Basically, OpenAI wants to use frontier AI to lock its own doors before attackers find an opening.
OpenAI just launched its Grid Protectors Program, agreeing to dial back power use at future AI compute campuses during peak grid strain. Its first pilot in Georgia can free up to 1,000 MW of demand response—enough to power 700,000 homes—while promising OpenAI will foot its own infrastructure bills so residents don't pay the tab.
Researchers introduced Weight-Space Programming, showing OpenAI's GPT-6 Astra can natively interpret and tweak neural network weights without traditional gradient training. By altering just 0.016% of Qwen3.5-0.8B's weights, it fixed a reasoning loop and bumped math accuracy from 7.9% to 48.1%. Why it matters: Direct parameter steering hints at automated model patching, though current results rely heavily on external Python tools rather than pure weight intuition.
OpenAI just revealed that starting in December, ChatGPT and Codex will share a single token allowance pool across Plus, Pro, and Business plans. The shift comes as OpenAI brings Work's agentic tools directly into regular chat, meaning heavier models and complex tasks will draw from the same cap, though instant chat stays untouched. If you burn through your quota, you will have to wait for resets or buy extra credits.
Google just agreed to adopt the IETF AIPREF standard alongside OpenAI, Anthropic, and Bing. Instead of blocking every AI crawler one by one, website owners will soon be able to use a shared policy token in robots.txt to opt out of training across all participating providers. Google plans to roll out support in early 2027.
OpenAI just published amicus briefs from the U.S. government, library groups, and legal scholars backing its defense against The New York Times. The filings argue AI training extracts statistical patterns rather than replacing original works, and curbing fair use would severely threaten scientific progress.
OpenAI just revealed that its unreleased internal models made major breakthroughs across math, computer science, and physics—including disproving the decades-old Exponential Time Hypothesis and advancing a Millennium Prize problem. The company published raw outputs and Lean formalizations on GitHub while recruiting top academics, including Nobel laureate David Gross, to steward open verification. If these proofs hold, AI has officially crossed the threshold into frontier scientific discovery.
OpenAI just launched GPT-6.1 Terra, a faster and cheaper model built specifically for visual reasoning and desktop computer use. It cuts standard API pricing by 50% down to $1 per million input tokens, while running tasks over three times faster than GPT-6 Astra. Basically, automating routine office software and technical drawings just became practical at scale.
OpenAI briefly listed two new release-candidate models on its API before pulling them down minutes later, exposing a dual-VM setup built for autonomous agents. A benchmark on OpenAI's blog also revealed a mystery internal model hitting 87.5% on engineering tasks, basically signaling the next flagship is right around the corner.
ByteDance’s video editor CapCut just announced a ChatGPT plugin that turns raw footage into a rough first cut through simple prompts. Creators can keep refining the timeline in chat—tweaking shot lengths and subtitles—or adjust things manually. It’s set to roll out for testing in the coming months.
