Recent News
Stories
A salvage yard in Nevada
How a family-run scrapyard used AI to catalogue 40,000 parts and cut customer wait time by 60%.
A seed farm in South Carolina
A fourth-generation farmer turns to AI to predict yields and reduce water usage across 800 acres.
A tamale shop in California
How one woman used ChatGPT to scale her grandmother's recipe into a regional business.
GPT-4 Technical Report
A large multimodal model accepting image and text inputs, emitting text outputs that exhibit human-level performance on various professional and academic benchmarks.
Scalable Oversight
Approaches to supervising AI systems that are more capable than their supervisors.
DALL·E 3
Understanding and generating images of unprecedented accuracy and detail.
Reinforcement Learning from Human Feedback
Training language models to follow instructions with human feedback to improve alignment.
We publish our research to advance the field and enable the broader community to build on our work.
All publicationsOur Models
Meet the
Intelligence
GPT-4o
Our most capable multimodal model. Processes text, images, and audio natively with human-level reasoning and sub-300ms response latency.
o3
OpenAI's most powerful reasoning model. Uses extended thinking to solve frontier-level math, science, and coding challenges step by step.
GPT-4o mini
The fastest, most cost-efficient model in the GPT-4o family. Ideal for high-volume tasks that need intelligence at scale without sacrificing quality.
DALL·E 3
The most accurate text-to-image model ever built. Understands nuanced prompts, generates photorealistic or artistic images with precise detail.
Trusted by leading organizations
From startups to Fortune 500s, teams worldwide use OpenAI to build products that matter.
Explore customer storiesAs featured in
The path to transformative AI requires confronting the hardest problems in safety head-on. We believe the best way to navigate these challenges is through open, rigorous research.
Alignment
Research to ensure AI systems act in accordance with human intentions and values.
Interpretability
Making AI systems understandable, so humans can verify their reasoning.
Robustness
Ensuring AI systems behave reliably across diverse conditions and adversarial settings.
BUILD WITH
INTELLIGENCE
Deploy intelligence that scales with your ambition. Start building today.