No-code platform opens early access to generative binder design with breakthrough hit rates and picomolar affinities

AI drug discovery startup Latent Labs has launched Latent-X, a new generative model designed to create functional, high-affinity protein binders at atomic resolution — without the need for coding or AI infrastructure.

Available now via early access on the company’s web-based no-code platform, Latent-X allows users to upload target proteins, generate macrocycles and mini-binders, and computationally score designs within minutes. The platform includes structure overlays, hotspot selection, and ranking tools to streamline design and prioritization — opening new opportunities for both scientists and non-AI specialists.

Latent Labs says its approach dramatically outperforms previous generative tools, delivering 91–100% hit rates for macrocycles and 10–64% for mini-binders across seven therapeutic targets in lab testing. In several cases, picomolar binding affinities were achieved — a performance that surpasses current generative benchmarks under identical conditions.

“We envision a future where effective therapeutics can be designed entirely in a computer,” said Simon Kohl, CEO and founder of Latent Labs. “Our platform empowers scientists with lab-validated protein binder design at their fingertips, whether they’re experts or new to AI-powered drug design. This is the first step in making biology programmable.”

Latent-X is designed to go beyond prediction to full de novo design. The model generates all-atom structures for previously untargeted proteins — solving binding interactions as geometric problems at the atomic level. Macrocycles, known for their potential oral deliverability and tissue penetration, and mini-binders, valued for their flexibility and specificity, are generated from scratch based on user inputs.

The company’s technical report details head-to-head comparisons against other models, where Latent-X consistently delivered fewer candidates with stronger affinity and higher success rates. According to Latent Labs, the model samples structure and sequence simultaneously, generating outputs over 10 times faster than earlier approaches — enabling experimentation and iteration within seconds.

With a free tier available for both commercial and academic users, the company aims to democratize access to AI protein design. Latent Labs is also seeking partnerships to expand use cases across therapeutic modalities, including nanobody and antibody development.

Backed by a $50M funding round announced earlier this year, Latent Labs brings together talent from AlphaFold 2, DeepMind, Apple, Exscientia, Mammoth Bio, and others. Investors include Radical Ventures, Sofinnova Partners, and high-profile figures such as Google’s Jeff Dean, Anthropic’s Dario Amodei, and Eleven Labs’ Mati Staniszewski.