# ArcInstitute/evo2

Genome modeling and design across all domains of life

Repository: https://github.com/ArcInstitute/evo2
Canonical: https://ross.abutalabs.com/products/evo2
Language: Jupyter Notebook
License: Apache-2.0
License Family: permissive
Last push: 2026-06-19T01:27:27+00:00

## Health v2 (maintenance only)
Score: 62/100 (v2, computed 2026-09-03T02:39:23.370411+00:00)
- activity 88, release rhythm 40, longevity 40
- inputs: {"age_days": 566, "days_push": 76, "days_rel": 187, "gap_med": null, "n_releases_24m": 1}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 4157, forks 548 (observed 2026-08-28T04:08:36.548866+00:00)

## What it is
Evo 2 is a DNA foundation language model (1B-40B parameters) that models genomes at single-nucleotide resolution with up to 1 million base pair context length, built on the StripedHyena 2 architecture. This repository provides inference, embedding, and generation code, plus training and finetuning support, for running Evo 2 locally or via Nvidia NIM.

## Use cases
- generate novel DNA sequences with a genome language model
- get embeddings for genomic sequences for downstream prediction
- predict variant effects from DNA sequences
- finetune a DNA foundation model on custom genomic data
- model long genomic contexts up to 1 million base pairs
- run a biological sequence model on GPU locally

## When to choose
- you need a state-of-the-art DNA/genome language model for inference or finetuning
- you work in genomics and want long-context sequence modeling at nucleotide resolution
- you have NVIDIA GPUs (Hopper for large models) and want local genome generation or embeddings

## When to avoid
- you need protein or RNA modeling rather than DNA
- you have no CUDA-capable NVIDIA GPU (large models require FP8 on Hopper)
- you want a lightweight CPU-only bioinformatics tool

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, deep-learning, llm-inference, llm-training, sdk
- domain: bioinformatics, machine-learning, deep-learning, large-language-models
- platform: python
- tags: dna-language-model, genomics, genome-design, long-context, stripedhyena, foundation-model, inference, linux, gpu, docker

## Member repositories
- ArcInstitute/evo2 (main) score 62

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:36.548866+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T18:22:51.010752+00:00, confidence not recorded.
  - readme: https://github.com/ArcInstitute/evo2 (fetched 2026-08-28T04:08:36.548866+00:00, sha 58ac1cf60187)
  - registry_pypi: https://pypi.org/pypi/evo2/json (fetched 2026-08-29T09:13:35.889947+00:00, sha 40c7062ef5b5)
- Data as of 2026-08-30T08:39:29.467469+00:00.