Chai Discovery has released Chai-1, a novel artificial intelligence model that rivals AlphaFold3 and pushes the boundaries of molecular structure prediction. This multimodal foundational model has achieved state-of-the-art performance in critical tasks such as drug discovery and biological research. Chai-1 is capable of predicting protein, small molecule, DNA, RNA, and other biological molecule structures. In multiple benchmark tests, it performs on par with or even outperforms leading models like AlphaFold 3. In the challenging PoseBusters protein-ligand benchmark test, Chai-1 achieves a success rate of 77%, comparable to AlphaFold 3's 76%. One key innovation of Chai-1 is its ability to generate accurate predictions from a single protein sequence without the need for multiple sequence alignment (MSA). In single-sequence mode, it surpasses existing MSA-dependent models in tasks such as predicting protein oligomer structures. The model is also capable of integrating experimental constraints, such as data from epitope mapping or cross-linking mass spectrometry. With the incorporation of these constraints, Chai-1's performance significantly improves, enhancing applications like antibody engineering where even minimal contact data can double the accuracy of antibody-antigen structure prediction. Researchers can use new data prompts in real-time to optimize predictions, making it a highly adaptable tool in complex biochemical research. Chai Discovery offers Chai-1 for free academic and commercial use through a web interface. Additionally, the model's weights and inference code are released as an open-source software library for non-commercial applications. This open release of the cutting-edge model is particularly noteworthy, especially considering DeepMind's shift towards a closed-source approach for AlphaFold 3 and the establishment of Alphabet's Isomorphic Labs, which focuses on technology monetization. Researchers and developers can access Chai-1 through the company's website or download the library from GitHub to integrate it into their own workflows.