@RadicalNumerics

Systems, scaling and architecture for general biological intelligence

 San Francisco & Tokyo
radicalnumerics.ai/
Joined May 2025
  • Tweets62
  • Following4
  • Followers4.4K
42Media
1/10 Introducing Omnii Genome, one of eight Advances built on Omnii-1. A single generalist model now outperforms leading specialized models (AlphaGenome) in linking genetic changes to gene activity and mapping regulatory interactions. These capabilities help make sense of the 98% of the human genome that doesn’t code for proteins How Omnii-1 works:
9
40
13
146
55,876
more replies
10/ Omnii Genome brings us closer to understanding how changes in DNA affect biological function, and eventually designing changes that produce desired effects. Explore the results: radicalnumerics.ai/blog/omni… Join us: radicalnumerics.ai/join-us
1
1
6
488
More thoughts from our own @m_m_carter xitter.cf/m_m_carter/status/2113…
Matt Carter@m_m_carterOct 22
Super excited about this release. One of the key advantages Omnii-1 has is its flexibility in accommodating new data. We explore some vignettes in our Omnii Genome blog post, and it’s also been a lot of fun pointing Omnii-1 at totally new tasks such as cancer diagnostics. (1/4)
2
332
6/ Omnii-1 also predicts DNA methylation at individual CpG sites with high accuracy. AlphaGenome, Borzoi, and ChromBPNet don’t predict DNA methylation.
1
5
432
7/ On identifying variants that affect gene expression (eQTLs), Omnii-1 reaches 0.810 AUROC. AlphaGenome reports 0.799 with a supervised predictor and 0.713 zero-shot.
1
5
399
4/ Omnii-1 uses DNA and available measurements from a biosample to infill missing assays. The predictions are specific to the supplied cellular context, so the same DNA can yield different reconstructions across cells and tissues.
1
1
5
485
5/ Here, we show Omnii-1 correctly predicts high expression in a liver-specific gene and no expression in skeletal muscle. Attribution arcs indicate which DNA regions and measured signals each predicted base is sensitive to, highlighting the promoter and exons.
1
6
448
Launching Omnii-1, our most powerful genome language model yet.
Eric Nguyen
@exnxOct 20
Introducing Omnii-1, our most powerful genome language model, and a major step toward general biological intelligence In a world first, Omnii-1 designed the first AI-generated bacteriophages to be administered to a human patient with an antibiotic-resistant infection under compassionate use
4
16
2
87
5,473
Radical Numerics retweeted
Eric Nguyen
@exnxOct 19
I'm thrilled to welcome Jill Hruby @jill_hruby and Paul Friedrichs as biodefense advisors to @RadicalNumerics! Jill formerly led the National Nuclear Security Administration, and Paul served as the inaugural White House Director of Pandemic Preparedness and Response. As AI opens new frontiers in biology, strengthening our ability to defend against biological threats is more important than ever. We're incredibly fortunate to learn from their experience and leadership as we advance this mission. Welcome aboard, Jill and Paul! Read the announcement: radicalnumerics.ai/blog/biod…

Radical Numerics Welcomes Jill Hruby and Paul Friedrichs as Biodefense Advisors

Jill Hruby and Paul Friedrichs join Radical Numerics as biodefense advisors, bringing deep experience in national security and biological risk reduction.

radicalnumerics.ai
5
1
25
2,634
1/8 Does text-to-image generation need language pretraining? Today, we share Halcyon, a family of early-fusion, base models trained from scratch. Halcyon outperforms strong late-fusion baselines with up to 275× less pretraining compute. Models and recipes are open. Halcyon is part of a research program at Radical Numerics on the science of multimodal pretraining and scaling.
4
31
3
178
75,767
more replies
8/8 Classifier-free guidance needs an unconditional image branch, usually trained by dropping captions. Halcyon instead places the caption after the image. The same model can then caption images at no extra pretraining cost, while keeping competitive generation quality.
1
5
914
read more: radicalnumerics.ai/blog/halc… paper: radicalnumerics.ai/assets/ha… website: cs.stanford.edu/~keshik/halc… We're hiring: radicalnumerics.ai/join-us Work done by @keshigeyan at @RadicalNumerics, @StanfordAILab and @StanfordSVL, with @__yuezhao__, @athmsx, @eadeli, @jcniebles, @jiajunwu_cs, @drfeifei and @MichaelPoli6.

Text-to-image generative modeling without language pretraining — Radical Numerics

Halcyon is a family of early-fusion multimodal autoregressive models pretrained from scratch on image-text pairs, with no internet-scale language pretraining.

radicalnumerics.ai
1
1
11
1,003
Load more