Profluent

Protein language models, AI-designed gene editors

USprivateprofluent.bio (opens profluent.bio)Checked 11 Oct

The brief

Analyst view
  • ProGen3 showed protein models get better with scale, like chatbots: bigger models make more varied working proteins.
  • Eli Lilly deal (April 2026) worth up to $2.25B in milestones to design recombinases that insert whole genes.
  • Raised $106M Series B (Altimeter, Bezos Expeditions) in late 2025; about $150M in total.

Technical approach

As reported
Data

Billions of natural proteins

Its Protein Atlas holds over 115 billion unique protein sequences mined from genomes, the raw text its models learn from.

Models

Language models for proteins

ProGen models learn the grammar of protein sequences and write new ones, as chat models write sentences.

Editors

Gene editors first

It designs CRISPR nucleases, base editors and recombinases, then licenses them or co-develops medicines with partners.

Primer