New Release
Inside our approach to the Model Spec
OpenAI unveils its Model Spec framework to enhance AI safety, user freedom, and accountability in model development.
OpenAI NewsOriginal →
News
2,420 stories
New Release
Introducing the OpenAI Safety Bug Bounty program
OpenAI launches a Safety Bug Bounty program to tackle AI abuse and enhance safety measures.
OpenAI NewsOriginal →
New Release
Helping developers build safer AI experiences for teens
OpenAI launches prompt-based teen safety policies to enhance AI experiences for teenagers.
OpenAI NewsOriginal →
New Release
Update on the OpenAI Foundation
OpenAI Foundation pledges $1 billion to initiatives targeting health, economy, and community support through AI advancements.
OpenAI NewsOriginal →
New Release
Powering product discovery in ChatGPT
ChatGPT introduces a new feature for immersive product discovery using the Agentic Commerce Protocol.
OpenAI NewsOriginal →
New Release
A New Framework for Evaluating Voice Agents (EVA)
Hugging Face unveils EVA, a new framework designed to standardize the evaluation of voice agents.
Hugging Face BlogOriginal →
New Release
Creating with Sora Safely
OpenAI launches Sora 2, prioritizing user safety in advanced video creation and social platforms.
OpenAI NewsOriginal →
New Release
Build a Domain-Specific Embedding Model in Under a Day
Create specialized embedding models quickly with Hugging Face's new approach, boosting productivity for AI developers.
Hugging Face BlogOriginal →
New Release
How we monitor internal coding agents for misalignment
OpenAI reveals its strategies for ensuring coding agents align with safety protocols through advanced monitoring techniques.
OpenAI NewsOriginal →
New Release
OpenAI to acquire Astral
OpenAI's acquisition of Astral is set to bolster Codex and enhance Python development tools.
OpenAI NewsOriginal →
New Release
State of Open Source on Hugging Face: Spring 2026
Hugging Face's Spring 2026 report reveals significant growth in open source AI models and community contributions.
Hugging Face BlogOriginal →
Developing
Measuring progress toward AGI: A cognitive framework
Google DeepMind introduces a new framework to measure advancements in artificial general intelligence through a Kaggle hackathon.
Google DeepMind BlogOriginal →
New Release
Measuring progress toward AGI: A cognitive framework
Google DeepMind unveils a new framework to measure progress toward AGI alongside a Kaggle hackathon for evaluation development.
Google DeepMind BlogOriginal →
New Release
Holotron-12B - High Throughput Computer Use Agent
Holotron-12B promises to revolutionize AI tasks with unprecedented computational efficiency and scalability.
Hugging Face BlogOriginal →
New Release
OpenAI Japan announces Japan Teen Safety Blueprint to put teen safety first
OpenAI Japan launches a comprehensive safety initiative aimed at protecting teenagers in the realm of generative AI.
OpenAI NewsOriginal →
New Release
Introducing GPT-5.4 mini and nano
OpenAI unveils GPT-5.4 mini and nano, optimized for speed and efficiency in coding and multimodal tasks.
OpenAI NewsOriginal →
New Release
Equipping workers with insights about compensation
Americans are turning to ChatGPT for compensation insights, sending nearly 3 million messages daily to bridge the wage information gap.
OpenAI NewsOriginal →
New Release
Why Codex Security Doesn’t Include a SAST Report
OpenAI's Codex Security opts for AI-driven methods over traditional SAST to enhance vulnerability detection.
OpenAI NewsOriginal →
New Release
Designing AI agents to resist prompt injection
OpenAI unveils new strategies for ChatGPT to combat prompt injection and enhance data security.
OpenAI NewsOriginal →
New Release
From model to agent: Equipping the Responses API with a computer environment
OpenAI introduces a new agent runtime that enhances the Responses API for scalable and secure file management.
OpenAI NewsOriginal →
New Release
Wayfair boosts catalog accuracy and support speed with OpenAI
Wayfair integrates OpenAI models to enhance product catalog accuracy and customer support efficiency.
OpenAI NewsOriginal →
New Release
Rakuten fixes issues twice as fast with Codex
Rakuten doubles its issue resolution speed by leveraging OpenAI's Codex technology.
OpenAI NewsOriginal →
Developing
Improving instruction hierarchy in frontier LLMs
OpenAI launches the IH-Challenge to enhance instruction hierarchy in large language models for better safety and steerability.
OpenAI NewsOriginal →