One year since GPT-4 deployment: From GPT-1 and 2 establishing the language model paradigm, through GPT-3's scaling predictions, to GPT-4 showing how complex systems emerge, mimicking nature’s unpredictable patterns from simple elements. An exploration from observation to deep,

Our early findings from an initial evaluation of Voice Engine, a model that generates speech closely resembling the source speaker's voice from text input and a 15-second audio sample. openai.com/blog/navigatin…

We just released improvements to GPT-4 Turbo

We've been tinkering to make our models smarter

Latest GPT-4 Turbo also out in ChatGPT now

Bringing memory to Plus users for a more personalized experience in ChatGPT

Good article on the work that Charles Fisher & his team at Unlearn.AI are doing, using deep learning to make clinical trials smarter — creating digital twins to predict how diseases progress, which speeds up the whole process and helps more people wsj.com/articles/the-a…

Very happy to have Kevin & Sarah on our team!

As AI improves humans will need more and more help to monitor and control it. So my team at OpenAI have trained an AI that helps humans to evaluate AI! (1/5)

GPT-4o mini makes intelligence far more affordable opening up a wide range of applications. Available on API and rolling out on ChatGPT today

You can now fine-tune GPT-4o mini. Enjoy!

Cranking up the accuracy and safety of our models, while cutting down on the need for a lot of human-labeled data.

Our SearchGPT prototype is now live. We’re looking for feedback as we prepare to integrate the experience in ChatGPT

Today we’re rolling out advanced Voice Mode in ChatGPT to small group of Plus users, all Plus users in fall. Richer and more natural real-time conversations make the technology less rigid — we’ve found it more collaborative and helpful and think you will as well.

Consider joining Karina Nguyen’s team to align and craft our models

Shared our GPT-4o system card today, with details of our extensive safety evaluations and mitigations implemented before deploying the model. We talk about specific risks, such as unauthorized voice generation and ungrounded inference, and how these were addressed through robust