How do I get access to AIDO Cell?

AIDO Cell is in early access. GenBio AI runs a sign-up list for scientists in academia and the biopharmaceutical industry, and says a more capable version with more cell types is planned for later in 2026. The 1.0 announcement points people to that list rather than to a public download or API.

Which cells can AIDO Cell 1.0 simulate?

AIDO Cell 1.0 ships prototype virtual cells for two human lines: K-562, which came from a leukemia patient, and Hep-G2, which came from a liver cancer patient. GenBio AI picked them because decades of multi-omics data exist to check the simulation against. A second system, AIDO Foundry, is meant to adapt the engine to other cell lines and primary cells.

How does AIDO Cell compare with AlphaFold 3 and AlphaGenome?

AIDO Cell 1.0 covers all five task families in GenBio AI's Virtual Cell Benchmark 1.0, while AlphaFold 3, Boltz-1, AlphaGenome, scGPT and Geneformer each address one or two. GenBio AI reports AIDO Cell leads on 24 of the 31 metrics, with AlphaFold 3 still ahead on protein-DNA interactions and AlphaGenome ahead on some genome-regulation tracks.

What can AIDO Cell do for drug design?

AIDO Cell supports what GenBio AI calls in-context molecular design: you describe the cell state you want instead of a binding pocket, and the system generates molecules, applies them to cloned cells and profiles each one. In an ABL1 test against the cancer drug imatinib, several designs with different chemistry recovered 65 to 88 percent of imatinib's gene-expression changes.

GenBio AI · 2026-08-18 · major

AIDO Cell — GenBio AI's virtual cell simulates drugs on a whole human cell

AIDO Cell 1.0 is a virtual-cell world model from GenBio AI that simulates a human cell from DNA and RNA up to whole-cell shape. GenBio AI reports state of the art on 24 of 31 metrics in its new Virtual Cell Benchmark 1.0.

AIDO Cell announcement graphic from GenBio AI

A world model that holds one cell state you can perturb, clone and read out across DNA, RNA, protein and cell shape.

Key specs

Benchmark sota	24 of 31 metrics
Simulated perturbations	1M per cell line

Quick facts

Maker	GenBio AI (Palo Alto)
Version	AIDO Cell 1.0
Cell lines	K-562 and Hep-G2
Benchmark	Virtual Cell Benchmark 1.0, 5 task families
Compared with	AlphaFold 3, Boltz-1, AlphaGenome, scGPT, Geneformer
Also ships	AIDO Foundry and AIDO Lab
Availability	Early access sign-up list

What is it?

AIDO Cell 1.0 keeps a single simulated cell state that you change step by step, instead of making one-shot predictions. GenBio AI published it on 18 August 2026 as a general-purpose simulator for two human cell lines. A researcher applies a gene knockout or a drug, then reads expression, protein structure, localization or cell shape out of that same state.

How does it work?

The simulation engine is a world model: every readout decodes from one shared internal state, so chromatin accessibility, expression, protein structure and morphology all describe the same cell at the same moment. A harness wraps the engine, tracking the cell across consecutive interventions and turning user commands into steps the engine can run. Cells can be cloned to branch an experiment, so a resistance mutation and two follow-up treatments can be compared from a shared ancestor.

Why does it matter?

Drug teams normally stitch together separate models — one for structure, one for expression, one for genome regulation — and the pieces do not agree with each other. GenBio AI reports that AIDO Cell 1.0 is the only system covering all five families in its Virtual Cell Benchmark 1.0, and that the simulation also runs backwards: name the cell state you want and it designs molecules to reach it. The company is a computational-biology startup whose co-founders include David Baker, who shared the 2024 Nobel Prize in Chemistry.

Who is it for?

computational biologists and drug discovery teams

Frequently asked questions

How do I get access to AIDO Cell?: AIDO Cell is in early access. GenBio AI runs a sign-up list for scientists in academia and the biopharmaceutical industry, and says a more capable version with more cell types is planned for later in 2026. The 1.0 announcement points people to that list rather than to a public download or API.
Which cells can AIDO Cell 1.0 simulate?: AIDO Cell 1.0 ships prototype virtual cells for two human lines: K-562, which came from a leukemia patient, and Hep-G2, which came from a liver cancer patient. GenBio AI picked them because decades of multi-omics data exist to check the simulation against. A second system, AIDO Foundry, is meant to adapt the engine to other cell lines and primary cells.
How does AIDO Cell compare with AlphaFold 3 and AlphaGenome?: AIDO Cell 1.0 covers all five task families in GenBio AI's Virtual Cell Benchmark 1.0, while AlphaFold 3, Boltz-1, AlphaGenome, scGPT and Geneformer each address one or two. GenBio AI reports AIDO Cell leads on 24 of the 31 metrics, with AlphaFold 3 still ahead on protein-DNA interactions and AlphaGenome ahead on some genome-regulation tracks.
What can AIDO Cell do for drug design?: AIDO Cell supports what GenBio AI calls in-context molecular design: you describe the cell state you want instead of a binding pocket, and the system generates molecules, applies them to cloned cells and profiles each one. In an ABL1 test against the cancer drug imatinib, several designs with different chemistry recovered 65 to 88 percent of imatinib's gene-expression changes.

Try it

Request early access at https://genbio.ai/waitlist