EvolutionaryScale logo

EvolutionaryScale

EvolutionaryScale is a New York-based biology foundation model lab spun out of Meta AI's ESM team that develops AI to deepen scientific understanding of biology. Its flagship ESM3 model is a multimodal generative protein language model that reasons jointly across sequence, structure, and function, scaling to 98B parameters trained on 771B tokens from 2.78B natural proteins. The companion ESM Cambrian (ESM C) family provides protein representation learning at 300M–6B parameters as a performant ESM2 replacement. Models are accessible via the hosted Forge inference API (forge.evolutionaryscale.ai), an open-source Python SDK (`pip install esm`), open weights on Hugging Face, and AWS Marketplace (SageMaker, NVIDIA BioNeMo and NIM). EvolutionaryScale was integrated into the Biohub organization in 2025; the ESM SDK now lives at github.com/Biohub/esm.

4 APIs 0 Features
AIArtificial IntelligenceBiologyBioinformaticsComputational BiologyDrug DiscoveryESMESM3ESM CambrianFoundation ModelsGenerative BiologyLife SciencesMachine LearningProtein DesignProtein FoldingProtein Language ModelsProteinsRepresentation LearningStructure Prediction

APIs

EvolutionaryScale Forge ESM3 API

Hosted inference API for the ESM3 multimodal protein language model. Reasons jointly across sequence, structure, and function tracks. Provides generate, batch_generate, encode, ...

EvolutionaryScale Forge ESM Cambrian API

Hosted inference API for the ESM Cambrian (ESM C) protein representation learning model family. Drop-in replacement for ESM2 offering comparable accuracy at lower memory footpri...

EvolutionaryScale Forge Folding API

Hosted folding and inverse-folding inference endpoints. `fold` predicts protein backbone coordinates plus pLDDT/PTM confidence from an input sequence; `inverse_fold` designs can...

EvolutionaryScale ESM Python SDK

Official Python SDK packaging ESM3 and ESM Cambrian model loaders, the `ESMProtein` multi-track data model, generation/sampling configurations, structure tokenization utilities,...

Semantic Vocabularies

Evolutionaryscale Context

0 classes · 5 properties

JSON-LD

API Governance Rules

EvolutionaryScale API Rules

7 rules · 3 errors 4 warnings

SPECTRAL

Sources

Raw ↑ 
aid: evolutionaryscale
url: https://raw.githubusercontent.com/api-evangelist/evolutionaryscale/refs/heads/main/apis.yml
apis:
  - aid: evolutionaryscale:forge-esm3-api
    name: EvolutionaryScale Forge ESM3 API
    tags:
      - AI
      - Biology
      - Foundation Models
      - Proteins
      - ESM3
      - Generation
    humanURL: https://forge.evolutionaryscale.ai
    properties:
      - url: https://forge.evolutionaryscale.ai
        type: Documentation
      - url: https://github.com/Biohub/esm
        type: SourceCode
      - url: openapi/evolutionaryscale-forge-esm3-api-openapi.yml
        type: OpenAPI
      - url: json-schema/evolutionaryscale-esm-protein-schema.json
        type: JSONSchema
      - url: json-schema/evolutionaryscale-generation-config-schema.json
        type: JSONSchema
      - url: json-ld/evolutionaryscale-context.jsonld
        type: JSONLD
    description: >-
      Hosted inference API for the ESM3 multimodal protein language model. Reasons jointly across sequence, structure,
      and function tracks. Provides generate, batch_generate, encode, decode, forward_and_sample, and logits operations
      across small (1.4B), medium (7B), and large (98B) parameter checkpoints. Accessed via the `esm` Python SDK (`pip
      install esm`) using a bearer token issued by forge.evolutionaryscale.ai. Closed beta with commercial license
      tiers.
  - aid: evolutionaryscale:forge-esmc-api
    name: EvolutionaryScale Forge ESM Cambrian API
    tags:
      - AI
      - Biology
      - Foundation Models
      - Proteins
      - ESM Cambrian
      - Embeddings
      - Representation Learning
    humanURL: https://forge.evolutionaryscale.ai
    properties:
      - url: https://forge.evolutionaryscale.ai
        type: Documentation
      - url: https://github.com/Biohub/esm
        type: SourceCode
      - url: openapi/evolutionaryscale-forge-esmc-api-openapi.yml
        type: OpenAPI
      - url: json-schema/evolutionaryscale-logits-output-schema.json
        type: JSONSchema
    description: >-
      Hosted inference API for the ESM Cambrian (ESM C) protein representation learning model family. Drop-in
      replacement for ESM2 offering comparable accuracy at lower memory footprint. Available in 300M, 600M, and 6B
      parameter sizes. Exposes encode and logits operations for generating protein sequence embeddings, hidden states,
      and per-residue logits for downstream representation tasks.
  - aid: evolutionaryscale:forge-folding-api
    name: EvolutionaryScale Forge Folding API
    tags:
      - AI
      - Biology
      - Foundation Models
      - Proteins
      - Structure Prediction
      - Inverse Folding
    humanURL: https://forge.evolutionaryscale.ai
    properties:
      - url: https://forge.evolutionaryscale.ai
        type: Documentation
      - url: https://github.com/Biohub/esm
        type: SourceCode
      - url: openapi/evolutionaryscale-forge-folding-api-openapi.yml
        type: OpenAPI
    description: >-
      Hosted folding and inverse-folding inference endpoints. `fold` predicts protein backbone coordinates plus
      pLDDT/PTM confidence from an input sequence; `inverse_fold` designs candidate sequences consistent with an input
      structure. Includes an `msa` endpoint for fetching multiple sequence alignments used to condition predictions.
  - aid: evolutionaryscale:esm-python-sdk
    name: EvolutionaryScale ESM Python SDK
    tags:
      - AI
      - Biology
      - SDK
      - Python
      - Open Source
      - ESM3
      - ESM Cambrian
    humanURL: https://github.com/Biohub/esm
    properties:
      - url: https://github.com/Biohub/esm
        type: SourceCode
      - url: https://pypi.org/project/esm/
        type: SDK
      - url: https://huggingface.co/biohub/esm3-sm-open-v1
        type: Documentation
      - url: https://huggingface.co/biohub/esmc-300m-2024-12
        type: Documentation
      - url: https://huggingface.co/biohub/esmc-600m-2024-12
        type: Documentation
      - url: https://github.com/Biohub/esm/tree/main/cookbook
        type: CodeExamples
    description: >-
      Official Python SDK packaging ESM3 and ESM Cambrian model loaders, the `ESMProtein` multi-track data model,
      generation/sampling configurations, structure tokenization utilities, and a `forge.client()` factory that swaps
      local checkpoints for Forge-hosted inference without code changes. Installable from PyPI as `esm`. Mixed
      commercial / non-commercial licenses.
name: EvolutionaryScale
tags:
  - AI
  - Artificial Intelligence
  - Biology
  - Bioinformatics
  - Computational Biology
  - Drug Discovery
  - ESM
  - ESM3
  - ESM Cambrian
  - Foundation Models
  - Generative Biology
  - Life Sciences
  - Machine Learning
  - Protein Design
  - Protein Folding
  - Protein Language Models
  - Proteins
  - Representation Learning
  - Structure Prediction
commonProperties:
  - url: https://www.evolutionaryscale.ai
    type: Portal
  - url: https://forge.evolutionaryscale.ai
    name: EvolutionaryScale Forge
    type: SignUp
  - url: https://github.com/Biohub/esm
    name: ESM SDK on GitHub
    type: SourceCode
  - url: https://pypi.org/project/esm/
    name: esm package on PyPI
    type: SDK
  - url: https://huggingface.co/biohub
    name: Biohub on Hugging Face
    type: Documentation
  - url: https://huggingface.co/biohub/esm3-sm-open-v1
    name: ESM3-open (1.4B) on Hugging Face
    type: Models
  - url: https://huggingface.co/biohub/esmc-300m-2024-12
    name: ESM C 300M on Hugging Face
    type: Models
  - url: https://huggingface.co/biohub/esmc-600m-2024-12
    name: ESM C 600M on Hugging Face
    type: Models
  - url: https://github.com/Biohub/esm/tree/main/cookbook
    name: ESM Cookbook
    type: CodeExamples
  - url: https://github.com/Biohub/esm/tree/main/cookbook/tutorials
    name: ESM Tutorials
    type: Tutorials
  - url: https://www.science.org/doi/10.1126/science.ads0018
    name: ESM3 — Science paper (Hayes et al. 2025)
    type: Documentation
  - url: https://www.evolutionaryscale.ai/blog/esm3-release
    name: ESM3 release announcement
    type: Blog
  - url: https://www.evolutionaryscale.ai/blog/esm-cambrian
    name: ESM Cambrian announcement
    type: Blog
  - url: https://www.evolutionaryscale.ai/blog
    type: Blog
  - url: https://aws.amazon.com/marketplace/seller-profile?id=seller-iw2nbscescndm
    name: EvolutionaryScale on AWS Marketplace (SageMaker)
    type: Marketplace
  - url: https://github.com/evolutionaryscale/esm-sagemaker
    name: ESM on Amazon SageMaker examples
    type: CodeExamples
  - url: https://github.com/evolutionaryscale/esm-partner
    name: Partner integrations
    type: CodeExamples
  - url: https://www.evolutionaryscale.ai/policies/cambrian-inference-clickthrough-license-agreement
    name: Cambrian Inference Clickthrough License Agreement
    type: TermsOfService
  - url: https://responsiblebiodesign.ai
    name: Responsible Biodesign Framework
    type: Documentation
  - url: https://bit.ly/esm-slack
    name: ESM Community Slack
    type: Forum
  - url: https://github.com/evolutionaryscale
    type: GitHubOrganization
  - url: https://github.com/Biohub
    name: Biohub GitHub Organization (ESM home)
    type: GitHubOrganization
  - url: plans/evolutionaryscale-plans-pricing.yml
    type: Plans
  - url: rate-limits/evolutionaryscale-rate-limits.yml
    type: RateLimits
  - url: finops/evolutionaryscale-finops.yml
    type: FinOps
  - url: vocabulary/evolutionaryscale-vocabulary.yml
    type: Vocabulary
  - type: Models
    data:
      - name: esm3-large-2024-03
        parameters: 98B
        description: Largest ESM3 checkpoint, trained on 771B tokens from 2.78B natural proteins; 1e24 FLOPs.
      - name: esm3-medium-2024-08
        parameters: 7B
        description: Mid-size ESM3 checkpoint suitable for most Forge inference workloads.
      - name: esm3-small-2024-08
        parameters: 1.4B
        description: Small ESM3 checkpoint; open weights as esm3-sm-open-v1 (non-commercial use).
      - name: esm3-open
        parameters: 1.4B
        description: Open weights of esm3-small (biohub/esm3-sm-open-v1 on Hugging Face).
      - name: esmc-6b-2024-12
        parameters: 6B
        description: Largest ESM Cambrian representation model.
      - name: esmc-600m-2024-12
        parameters: 600M
        description: Mid-size ESM Cambrian representation model.
      - name: esmc-300m-2024-12
        parameters: 300M
        description: Small ESM Cambrian model; ESM2 650M-class quality with reduced memory footprint.
  - type: Features
    data:
      - ESM3 — multimodal generative model jointly conditioning on protein sequence, structure, and function
      - 98B-parameter ESM3 trained on 771B tokens from 2.78B natural proteins (1e24 FLOPs)
      - ESM Cambrian (ESM C) representation models at 300M, 600M, and 6B parameters
      - Forge API providing generate, batch_generate, encode, decode, forward_and_sample, and logits operations
      - Fold and inverse-fold endpoints for structure prediction and structure-conditioned sequence design
      - MSA endpoint for fetching multiple sequence alignments used by structure prediction
      - Iterative masked sampling with configurable num_steps, temperature, top_p, and decoding schedules
      - Per-track generation across sequence, structure, secondary_structure, sasa, and function tracks
      - Structure tokenizer converting PDB / atom37 coordinates to and from discrete tokens
      - ESMProtein and ESMProteinTensor data model unifying raw and tokenized representations
      - Async/sync client surface (`async_generate`, `async_fold`, `async_encode`, ...) for high-throughput jobs
      - Drop-in Forge client (`esm.sdk.client(model, token=...)`) replaces local checkpoints with hosted inference
      - Open-weights ESM3-open (1.4B) and ESM Cambrian distributions on Hugging Face under research license
      - AWS Marketplace deployment via SageMaker, NVIDIA BioNeMo, and NVIDIA NIM microservice
      - Cookbook tutorials covering protein generation, embedding workflows, and esmGFP-style design
      - Responsible Biodesign Framework governing model release and biosecurity review
    sources:
      - https://www.evolutionaryscale.ai
      - https://github.com/Biohub/esm
      - https://forge.evolutionaryscale.ai
      - https://www.science.org/doi/10.1126/science.ads0018
      - https://www.evolutionaryscale.ai/blog/esm3-release
      - https://www.evolutionaryscale.ai/blog/esm-cambrian
    updated: '2026-05-24'
created: '2026-05-24'
modified: '2026-05-24'
position: Consuming
description: >-
  EvolutionaryScale is a New York-based biology foundation model lab spun out of Meta AI's ESM team that develops AI to
  deepen scientific understanding of biology. Its flagship ESM3 model is a multimodal generative protein language model
  that reasons jointly across sequence, structure, and function, scaling to 98B parameters trained on 771B tokens from
  2.78B natural proteins. The companion ESM Cambrian (ESM C) family provides protein representation learning at 300M–6B
  parameters as a performant ESM2 replacement. Models are accessible via the hosted Forge inference API
  (forge.evolutionaryscale.ai), an open-source Python SDK (`pip install esm`), open weights on Hugging Face, and AWS
  Marketplace (SageMaker, NVIDIA BioNeMo and NIM). EvolutionaryScale was integrated into the Biohub organization in
  2025; the ESM SDK now lives at github.com/Biohub/esm.
maintainers:
  - FN: Kin Lane
    email: info@apievangelist.com
    X: apievangelist
    url: https://apievangelist.com
specificationVersion: '0.16'