Releases

OpenAI introduced DALL-E 2 with higher-resolution image generation and editing capabilities.

OpenAI introduced instruction-following models trained with human feedback.

OpenAI released a new endpoint and model families for text and code embeddings.

OpenAI introduced Codex, a descendant of GPT-3 that translates natural language into code.

OpenAI introduced CLIP, a neural network trained on image-text pairs for transferable visual recognition.

OpenAI introduced DALL-E, a model that generates images from text descriptions.

OpenAI launched a general-purpose API for accessing its language models.

OpenAI published GPT-3, demonstrating few-shot learning at 175 billion parameters.

OpenAI released Jukebox, a neural network that generated music including rudimentary singing as raw audio.

OpenAI completed the staged GPT-2 release with the 1.5-billion-parameter model and accompanying research.

OpenAI released the 774-million-parameter GPT-2 model and shared further analysis of misuse and detection.

Microsoft invested $1 billion in OpenAI and the companies formed an exclusive cloud-computing partnership.

OpenAI introduced MuseNet, a deep neural network for generating multi-instrument musical compositions.

OpenAI Five defeated the reigning Dota 2 world champion team OG in a live best-of-three match.

OpenAI created a capped-profit entity, OpenAI LP, to scale investment in compute and talent while keeping the nonprofit mission in control.

OpenAI introduced GPT-2 and began a staged release process due to misuse concerns.

OpenAI presented its five-agent Dota 2 system trained through large-scale self-play.

OpenAI published the original GPT approach: generative pre-training followed by discriminative fine-tuning.

OpenAI released code for training reinforcement-learning agents from human feedback.

OpenAI introduced PPO, a family of policy-gradient methods designed for stability and simplicity.

OpenAI and DeepMind researchers showed how agents can learn complex behavior from human preference comparisons.

OpenAI released Baselines, open-source high-quality implementations of reinforcement-learning algorithms.

OpenAI released Universe, a software platform for training intelligent agents across games, websites, and applications.

OpenAI released the public beta of Gym, a toolkit for developing and comparing reinforcement-learning algorithms.

OpenAI launched as a nonprofit artificial intelligence research company with a mission to ensure artificial general intelligence benefits all of humanity.