9/17/2024
Using GPT-4 to improve teaching and learning in Brazil
GPT-4 applied to education in Brazil; minimal technical details.
9/12/2024
Learning to reason with LLMs
OpenAI o1 uses reinforcement learning to perform complex reasoning with internal chain-of-thought before responding.
9/12/2024
Coding with OpenAI o1
Cognition CEO Scott Wu explains how OpenAI o1 improves coding by making human-like reasoning decisions.
9/12/2024
Decoding genetics with OpenAI o1
Geneticist Catherine Brownstein demonstrates OpenAI o1 diagnosing rare genetic medical conditions.
9/5/2024
Using GPT-4 to deliver a new customer service standard
Ada uses GPT-4 to improve customer service standards.
8/26/2024
Personalizing education with ChatGPT
Arizona State University deploys ChatGPT campus-wide for personalized learning and research.
8/20/2024
Fine-tuning now available for GPT-4o
OpenAI enables fine-tuning of GPT-4o models for custom applications and improved accuracy.
8/8/2024
GPT-4o System Card
GPT-4o system card documenting safety evaluations, red teaming, risk frameworks, and mitigations pre-release.
7/30/2024
A Primer on the EU AI Act: What It Means for AI Providers and Deployers
Overview of EU AI Act requirements and deadlines for AI providers and deployers, focusing on prohibited and high-risk uses.
7/24/2024
Improving Model Safety Behavior with Rule-Based Rewards
Rule-Based Rewards method aligns model behavior for safety without extensive human annotation during RLHF.
7/17/2024
Prover-Verifier Games improve legibility of language model outputs
Prover-verifier games framework improves interpretability and verifiability of language model outputs through structured reasoning.
6/27/2024
Finding GPT-4’s mistakes with GPT-4
CriticGPT, a GPT-4 based model, generates critiques of ChatGPT outputs to improve RLHF training and error detection.
6/20/2024
Improved Techniques for Training Consistency Models
Improved training techniques for consistency models enabling high-quality single-step generation without adversarial training.
6/20/2024
Consistency Models
Consistency models overcome slow iterative sampling in diffusion models, enabling fast one-step generation for images, audio, and video.
6/18/2024
Surging developer productivity with custom GPTs
Paf company achieved developer productivity gains by deploying ChatGPT Enterprise across engineering teams and coding academy.
6/17/2024
Using GPT-4o reasoning to transform cancer care
Color Health's Cancer Copilot uses GPT-4o to identify missing diagnostics and create treatment plans for cancer patients.
6/9/2024
Let's reproduce GPT-2 (124M)
Detailed reproduction of GPT-2 (124M parameter model) from scratch with training code and methodology.
6/7/2024
Expanding on how Voice Engine works and our safety research
OpenAI describes technology and safety research behind Voice Engine text-to-speech model.
6/6/2024
Extracting Concepts from GPT-4
New techniques for scaling sparse autoencoders identified 16 million interpretable patterns in GPT-4's internal computations.
5/29/2024
Automating customer support agents
MavenAGI launches GPT-4 based customer service agent used by Tripadvisor, Clickup, Rho to automate support.
5/16/2024
Improvements to data analysis in ChatGPT
ChatGPT data analysis improved with table/chart interaction and direct file integration from Google Drive and OneDrive.
5/13/2024
Hello GPT-4o
GPT-4 Omni flagship model supports real-time reasoning across audio, vision, and text.
5/13/2024
Introducing GPT-4o and more tools to ChatGPT free users
GPT-4o and additional tools released to free ChatGPT users.
5/13/2024
Spring Update
OpenAI announces GPT-4o model and expanded free ChatGPT capabilities.
4/24/2024
GPT-4 API general availability and deprecation of older models in the Completions API
GPT-4 API becomes generally available; GPT-3.5 Turbo, DALL-E, and Whisper APIs released with model deprecation plan.
4/24/2024
Introducing ChatGPT and Whisper APIs
ChatGPT and Whisper model APIs released for developer integration into applications.
4/23/2024
Introducing more enterprise-grade features for API customers
OpenAI releases enterprise API features including Assistants API updates and cost management tools.
4/19/2024
The Instruction Hierarchy: Training LLMs to Prioritize Privileged Instructions
Research on training LLMs to resist prompt injections and jailbreaks through instruction hierarchy.
4/5/2024
Klarna's AI assistant does the work of 700 full-time agents
Klarna deploys AI assistant handling workload of 700 customer service agents.
4/4/2024
Introducing improvements to the fine-tuning API and expanding our custom models program
OpenAI releases fine-tuning API improvements and custom models program expansion.
3/27/2024
OpenAI’s comment to the NTIA on open model weights
OpenAI submits policy comment to NTIA on open model weight regulations.
3/18/2024
Enterprise-ready trust and safety
Salesforce integrates OpenAI enterprise LLMs to transform customer applications.
3/6/2024
Sparking a more productive company with ChatGPT Enterprise
Match Group uses ChatGPT Enterprise to improve creativity and business impact.
2/20/2024
Let's build the GPT Tokenizer
Tutorial implementing byte-pair encoding tokenizer used in GPT models, with code walkthrough.
2/15/2024
Video generation models as world simulators
OpenAI releases Sora, a video generation model trained on videos and images with transformer architecture on spacetime patches, generating minute-long high-fidelity video.
2/13/2024
Memory and new controls for ChatGPT
ChatGPT gains memory feature allowing conversations to persist across sessions with user controls.
1/31/2024
Building an early warning system for LLM-aided biological threat creation
OpenAI develops evaluation framework for LLM risks in biological threat creation; GPT-4 shows mild uplift, establishing baseline for research.
1/25/2024
New embedding models and API updates
OpenAI launches new embedding models, GPT-4 Turbo variants, moderation models, API management tools, and pricing updates.
1/16/2024
Democratic inputs to AI grant program: lessons learned and implementation plans
OpenAI funds 10 teams to design AI governance tools and collective decision-making mechanisms; shares learnings and calls for participation.
1/10/2024
Introducing ChatGPT Team
Digital Green uses OpenAI to help farmers increase income through agricultural database tools.
1/10/2024
Introducing the GPT Store
ChatGPT Team plan offers secure collaborative workspace for teams to use ChatGPT at work.
12/14/2023
Superalignment Fast Grants
Summer Health uses OpenAI to improve accuracy of pediatric medical visit documentation.
12/14/2023
Practices for Governing Agentic AI Systems
$10M grants for research on superhuman AI alignment, interpretability, and scalable oversight techniques.
11/23/2023
[1hr Talk] Intro to Large Language Models
1-hour educational talk introducing key concepts and history of large language models for beginners.
11/9/2023
OpenAI Data Partnerships
OpenAI partners to create open-source and private datasets for AI model training.
11/6/2023
Introducing GPTs
OpenAI launches custom ChatGPT versions combining instructions, knowledge, and skills for personalized AI assistants.
11/6/2023
New models and developer products announced at DevDay
GPT-4 Turbo with 128K context, Assistants API, Vision, and DALL·E 3 API announced at DevDay.
9/25/2023
ChatGPT can now see, hear, and speak
ChatGPT now supports voice conversation and image input capabilities.
9/25/2023
GPT-4V(ision) system card
GPT-4V(ision) system card detailing vision capabilities and safety testing.
9/19/2023
OpenAI Red Teaming Network
OpenAI Red Teaming Network seeks domain experts to improve model safety through testing.