About six months after getting theft with $ 50 million, Latent Labs has released an online-based AI model for programming biology.
The Latent Labs model has “achieved high -tech measurements” when testing proteins developed by a physical laboratory, according to high -tech or war, is often used in the AI field, which represents the best performance in the industry in a particular task.
“We have computational ways to judge how good the model
The company’s fundamental biology model, known as Lattx, enables academic institutions, companies starting biotechnology, and pharmaceutical companies to design new proteins directly into the browser using natural language.
Latentx exceeds what is found in nature by creating completely new molecular design, such as nanobodies and newcomers with accurate atomic structures. This approach can help to develop new therapeutic drugs much faster.
According to Kohl, this ability to design completely new proteins is what separates Latentx from the Alfafold.
“Alfa -Fold is a model for predicting a protein structure. So it allows you to visualize existing structures, but it does not, it doesn’t let you produce new proteins,” he said.
TechCrunch event
San Francisco
And
27.-29. October 2025
Contrary to AI-based drug discovery attempts, such as Xira, Recursion or Deepmind Spinout Isomorphic Laboratories, which focus on developing ownership medicines, Latent Labs’ business model includes a license for external organizations.
“Not every company has the opportunity to build their own AI models, get their own ai infrastructure and get their own AI teams,” Kohl said.
Although Latentx is available for free, Kohl said that the company will eventually charge for advanced features and features when presented.
Other companies that provide open AI’s founding models for finding drugs are Chai Discovery and Evolutionaryryscale.
Latent Labs supported radical Ventures, Sofinnova Partners, Google Chief Researcher Jeff Dean, Anthropis CEO Dario Amodei and Eleven Labs CEO Mati Staniszewski.
