Skip to content

Model Labs

Arc Institute

Independent research institute building genomic foundation models, including the Evo series.

DNA embeddingsSequence generationSequence scoringVariant effects4 models

Models on Rafflesia

Evo 2 7B Base

activeApache-2.0
arc-institute/evo2/7b-base/embed

Create pooled or per-token DNA representations with an explicit layer and model release.

DNA embeddingsdna_sequenceembedding_vector
$0.01 / 1K bases

Evo 2 7B Base

activeApache-2.0
arc-institute/evo2/7b-base/likelihood

Score a nucleotide sequence as an Evo 2 log-likelihood under the genome model, returning the total, the length-normalised mean, and optional per-position log probabilities.

Sequence scoringdna_sequencesequence_log_likelihood + position_log_probabilities
$0.01 / 1K bases scored

Evo 2 7B Base

activeApache-2.0
arc-institute/evo2/7b-base/generate-sequence

Generate nucleotide sequence with Evo 2, optionally continuing a supplied prompt, under explicit length, sampling, and seed controls.

Sequence generationdna_sequencedna_sequence + sequence_log_likelihood
$0.02 / 1K bases generated

Evo 2 7B Base

activeApache-2.0
arc-institute/evo2/7b-base/variant-effects

Score nucleotide substitutions and indels as Evo 2 delta log-likelihoods over an explicit reference window, with the window, strand policy, and reference likelihood returned alongside every score.

Variant effectsdna_sequencelog_likelihood_ratio + mean_log_probability + position_log_probabilities
$0.01 / 1K bases scored

About

The Arc Institute is an independent nonprofit research organization in Palo Alto that funds scientists to pursue long-horizon research, in partnership with Stanford, UC Berkeley, and UCSF. Its machine learning work includes Evo, a family of genomic foundation models trained across the tree of life.

On Rafflesia, Arc's Evo 2 7B Base DNA embedding model is served for production use via the runs API.

Use these models

Every endpoint speaks the same run lifecycle: POST the endpoint id and a typed input, get a run back with outputs, provenance, and measured billing.

run.sh
# one run lifecycle, every endpoint
curl https://api.rafflesia.ai/v1/runs \
  -H "Authorization: Bearer $RAFFLESIA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
        "endpoint": "arc-institute/evo2/7b-base/embed",
        "input": { "sequence": "MKTAYIAK..." }
      }'