The brief
- Evo 2 (Feb 2025, with NVIDIA) learned from DNA of all known life and can predict effects of mutations and write new sequences.
- In 2024 Arc researchers showed 'bridge RNAs' that guide enzymes to insert, delete or flip DNA.
- The second Virtual Cell Challenge launched on 20 Aug 2026.
Technical approach
Language models for DNA
Evo reads DNA letters like words in a very long text, learning its grammar so it can predict mutation effects and design new sequences.
Predicting how cells react
Models trained on huge single-cell datasets predict how a cell's genes respond when a gene is switched off or a drug added.
Bridge editing
A 'bridge RNA' grabs both a target and a donor piece of DNA and guides an enzyme to rearrange them without cutting both strands.
Arc Institute, two programs
Updated 20 Aug 2026Research & demos
- Feb 2025paperGenome modeling and design across all domains of life with Evo 2 (preprint) (opens doi.org)Evo genomic foundation models
Scaled the model up and released it openly with NVIDIA.
- Nov 2024paperSequence modeling and design from molecular to genome scale with Evo (Science) (opens doi.org)Evo genomic foundation models
First Evo model: showed a DNA AI can design working CRISPR systems.
- Jun 2024paperBridge RNAs direct programmable recombination of target and donor DNA (Nature) (opens doi.org)Evo genomic foundation models
Introduced a new way to program DNA rearrangements.
Open technical problems
- Do DNA models understand function?Active research
AI-written DNA can look natural but fail in cells; slow lab testing limits how fast the models improve.
- Bridge editing in human cellsActive research
Bridge editing must work well and precisely in human cells before it can be a treatment.
Upcoming catalysts
- 22 Oct20262026 Virtual Cell Challenge: final test set releasedVirtual Cell Challengein 12 d
- Nov20262026 Virtual Cell Challenge winners announcedVirtual Cell Challengeby ~2 mo
Peers
Company & funding
- HQ
- US
- Founded
- 2021
Partners & customers
- Partner
Co-developed Evo 2 and sponsors the Virtual Cell Challenge.
- Partner
University partner alongside UC Berkeley and UCSF.
People
- Patrick Hsu (opens arcinstitute.org)Co-founder and core investigator
- Silvana Konermann (opens arcinstitute.org)Co-founder and executive director
- Patrick Collison (opens en.wikipedia.org)Co-founder

