Blog

We scale data mixing for language and genome models: thousands of sources derived from document behavior, and a Pareto frontier of mixtures over them.

Christine Parthemore joins Radical Numerics as Head of Biodefense, bringing two decades of experience at the U.S. Department of Defense and the Council on Strategic Risks.

We post-trained Omnii, our next-generation genome language model, for end-to-end cancer vaccine design, from neoantigen selection to the final mRNA.

Jacob Rinaldi joins Radical Numerics as Head of Scientific Strategy, bringing scientific depth from Genentech, Recursion, and Noetik to our work on general biological intelligence.

We've raised $50M to build general biological intelligence, and we're previewing Omnii, our next-generation genome language model for both human health and biodefense.

Omnii's first biodefense results, spanning paraphrased pathogen detection and fitness scoring of viral variants beyond standard methods.

A research preview of Omnii, our next-generation genome language model.

A deep dive into TransformerEngine's implementation of the NVFP4 recipe: nvfp4 data flow, custom kernels, and systems optimizations for performant FP4 training.

NVIDIA's NVFP4 pretraining recipe: floating point fundamentals, the evolution of mixed-precision training, and the techniques that enable stable NVFP4 training.

Phalanx: a drop-in replacement for sliding window attention. Faster and better quality. Hardware- and numerics-aware design of new layers that push the efficiency-quality frontier

Introducing RND1-30B, the largest open-source diffusion language model, trained via AR-to-Diffusion conversion with a simple, scalable recipe.