Hello, arXiv visitor 👋
Welcome to Replicate. We're making it easy to share your machine learning model. Check out some recently featured models:

deepseek-ai / deepseek-r1
A reasoning model trained with reinforcement learning, on par with OpenAI o1
355.4K runs

anthropic / claude-3.5-sonnet
Anthropic's most intelligent language model to date, with a 200K token context window and image understanding (claude-3-5-sonnet-20241022)
10.4K runs

black-forest-labs / flux-1.1-pro
Faster, better FLUX Pro. Text-to-image model with excellent image quality, prompt adherence, and output diversity.
16.1M runs
Share your model
If this is your first time creating a model, have a look at the docs, otherwise go ahead and create a new model.