Evo 2 (Claude Skill)

Generates and scores biological sequences across DNA, RNA, and protein with the Evo 2 genome language model, driven as a Claude skill.

   
Type Claude Skill
Supplier Arc Institute
Availability GA — Claude Science research skill
Pricing Free / OSS (Apache-2.0); NVIDIA-hosted inference metered per NVIDIA terms
Capabilities Read/Write — sequence generation and variant likelihood scoring
Verified works · 2026-07-20
Security cleared · 2026-07-20 — provenance matches ArcInstitute/evo2, Apache-2.0, active, no advisories

How to install

  • Claude Science — enable the built-in Evo 2 research skill (Anthropic-hosted; not published to the public anthropics/life-sciences marketplace).
  • Run the model yourself — the upstream model is open source:
    git clone https://github.com/ArcInstitute/evo2
    

    Follow the repo README for environment setup and model weights.

What it does

Runs Evo 2 for long-context genomic sequence modelling: de novo sequence generation, zero-shot variant-effect scoring, and embedding extraction. Local inference needs a Hopper-class GPU (FP8); otherwise use the NVIDIA-hosted inference API (NVIDIA NIM / build.nvidia.com).

Primary use cases: Genomic sequence generation, zero-shot variant-effect prediction, regulatory-element design

Notes

Claude Science: Featured as a research skill in Anthropic’s Claude Science. Its inclusion there is an independent signal of quality and trustworthiness for life-science work.

Local runs require a Hopper GPU with FP8 support; a hosted NVIDIA NIM endpoint is the alternative.

Sources


Installed this tool?

Share feedback — install path, OS, errors, workarounds. The form opens with this tool pre-selected and a link back to this page.