How Deutsche Telekom is rewiring telecommunications with AI

Inside the rollout Inside the rollout Results at a glance The next frontier is reinventing voice Leadership lessons Tips What's next Inside the rollout Results at a glance The next frontier is reinventing voice Leadership lessons Tips What's next Deutsche Telekom⁠(opens in a new

GPT-5.6: Frontier intelligence that scales with your ambition

Efficient by default, maximum performance on demand Efficient by default, maximum performance on demand A leap forward in design End-to-end knowledge work Pushing the frontier on cyber and science GPT-5.6 accelerates OpenAI Scaling safety and security with capability  Availability and prici

ChatGPT is now a partner for your most ambitious work

Work anywhere. Go further on desktop. Work anywhere. Go further on desktop. Create slides, sheets, docs, and Sites from your apps and workflows Delegate repetitive tasks to focus on work that matters Get work done faster across the web and your desktop apps From goals to real outcomes across

GPT-5.5 Bio Bug Bounty

Invitation Invitation How to participate Invitation How to participate Invitation As part of our ongoing efforts to strengthen our safeguards for advanced AI capabilities in biology, we’re evolving our GPT‑5.5 Bio Bug Bounty to become an ongoing private program—the OpenAI Bio Bounty

GPT-5.6 is now the preferred model in Microsoft 365 Copilot

Today, OpenAI announced GPT‑5.6, which will become the new preferred model in Microsoft 365 Copilot—in Word, Excel, PowerPoint, Chat and Cowork. For Microsoft 365 customers, the update brings OpenAI's latest flagship model series into productivity tools people use every day, helping them create, ana

Our approach to government and national security partnerships

Governments are beginning to use frontier AI systems for increasingly important work, including in national security. That creates new imperatives for AI labs, governments, and civil society to work together on how these tools should be used in sensitive settings.  We believe democratic societies s

Separating signal from noise in coding evaluations

Methodology Methodology Human-supervised agent review Human annotation campaign Discussion Methodology Human-supervised agent review Human annotation campaign Discussion Accurately measuring our models’ capabilities is important for sound deployment and safety decisions, including d

Helping K–12 educators build practical AI skills

Every day, K–12 teachers, administrators, district leaders, and school communities help students learn, grow, and prepare for the future. They are also being asked to navigate rapid technological change while managing limited time, growing responsibilities, and uneven access to practical support. T

Introducing GPT-Live

We’re launching GPT‑Live, a new generation of voice models that make talking with AI feel much more like having a real conversation. GPT‑Live is built on a full-duplex architecture, meaning it can listen and speak at the same time. During conversations, GPT‑Live can show it’s paying attention with

Australian Payments Plus moves faster with ChatGPT and Codex

Accelerating technical investigation Accelerating technical investigation Moving faster through payments complexity Turning rough inputs into decision-ready work Testing ideas earlier with ChatGPT and Codex Leadership lessons What's next Accelerating technical investigation Moving fast

How ChatGPT adoption has expanded

ChatGPT users deepened and expanded their use over time ChatGPT users deepened and expanded their use over time ChatGPT adoption accelerated across regions ChatGPT’s user base became more global and more diverse Learn more ChatGPT users deepened and expanded their use over time ChatGPT a

Core dump epidemiology: fixing an 18-year-old bug

First debugging attempt: carefully examining a few core dumps First debugging attempt: carefully examining a few core dumps Clues from the stack Doctor or epidemiologist? Cleaning the data Bug #1: the bad host Exception handling is a dynamic control transfer Undoing one last assumption Bug

Inside Genebench-Pro

Introducing GeneBench-Pro

Dataset construction Dataset construction Evaluation and grading Results Dataset construction Evaluation and grading Results Scientific data rarely arrive with instructions. Researchers must decide whether a pattern reflects biology or noise, whether the data can support the question

Mapping Europe’s AI Workforce Opportunity

AI capabilities can cross borders quickly. Jobs do not change in such a frictionless way. Work is shaped by licensing systems, local institutions, and the practical realities of delivering care, education, justice, public services, and other forms of human support. These systems matter because they

HP Inc. launches Frontier strategic partnership with OpenAI

From pilot wins to enterprise deployment From pilot wins to enterprise deployment Frontier as a connective layer Building an AI-driven operating model Table of contents From pilot wins to enterprise deployment Frontier as a connective layer Building an AI-driven operating model Enterp

Previewing GPT-5.6 Sol: a next-generation model

0 0 . 0 . . 0 . . . . 0 . 0 0 0 0 . 0 . . . 0

How agents are transforming work

Agentic AI changes the unit of knowledge work from single interactions to delegated, long-horizon tasks. Chatbot interactions are often short and self-contained. Agents can operate independently for minutes or hours while orchestrating tool calls, interacting with environments, and iterating towards

OpenAI and Broadcom unveil LLM-optimized inference chip

Designed to be the best inference platform for LLMs Designed to be the best inference platform for LLMs Nine-month tape-out, accelerated by OpenAI models Building a multi-generation platform with partners Making advanced AI more broadly available Table of contents Designed to be the best

How GPT-5 helped immunologist Derya Unutmaz solve a 3-year-old mystery

Solving a problem with GPT-5 Pro Solving a problem with GPT-5 Pro What this means for scientific research Table of contents Solving a problem with GPT-5 Pro What this means for scientific research Doctor and immunologist Derya Unutmaz has been interested in artificial intelligence for y

Helping build shared standards for advanced AI

Increasingly capable models can strengthen cyber defense, accelerate scientific discovery, and expand access to expertise. But they can also create safety and security risks if their capabilities are misunderstood, their safeguards are inadequate, or governments lack the information they need to res

How Omio is building the future of conversational travel

Moving from search to conversation Moving from search to conversation Bringing conversational travel to life Inside the rollout Results at a glance Leadership lessons What’s next Table of contents Moving from search to conversation Bringing conversational travel to life Inside the rol

Patch the Planet: a Daybreak initiative to support open source maintainers

How Patch the Planet works How Patch the Planet works Early field notes and findings from developers What OpenAI Daybreak is already finding Operating systems Network Browsers Shared infrastructure, shared defense Table of contents How Patch the Planet works Early field notes and fin

Daybreak: Tools for securing every organization in the world

Cyber defense at an inflection point Cyber defense at an inflection point From findings to fixes with Codex Security Updating GPT-5.5-Cyber: pairing capability with permissiveness Working with the security ecosystem Patch the Planet: landing fixes in open-source Protecting critical infrastru

Codex-maxxing for long-running work

Organizations are increasingly using AI to support work that extends beyond a single prompt. In this whitepaper, Jason Liu shares practical strategies for using Codex as a persistent workspace that preserves context, manages complex workflows, and helps sustain progress across long-running projects.