Loading…
Cloud catalog
Amazon Web Services
AWS Marketplace listings and SageMaker JumpStart deploy biological foundation models onto managed endpoints in the subscriber's own AWS account.
- Models served
- 11
- Offers
- InferenceFine-tuning
- Coverage
- Protein 4Small molecule 3Pathology 3DNA & Gene 1+3 more
- Deployment
- Managed cloud · your region
- Pricing
- Per token or per instance-hour
Models you can run on Amazon Web Services
Most cited first2K—1.8K
Message passing neural network for fixed-backbone protein sequence design. Achieves 52.4% native sequence recovery, far surpassing Rosetta's 32.9%.
Protein
571—4.2K
Open model that jointly predicts biomolecular structure and small-molecule binding affinity, approaching FEP+ accuracy in seconds on a single GPU.
Protein
3672.8K2.9K
Multimodal generative protein language model reasoning jointly over protein sequence, structure, and function, trained at 98B parameters.
Protein
3001.5K4.2K
Genomic foundation model trained on 9.3 trillion DNA base pairs across all domains of life, with 40B parameters and a 1-million-token context.
DNA & Gene
6—1
Generative chemistry foundation model pairing a text-and-SMILES language backbone with a 3D molecular point-cloud encoder for prediction and design.
Small moleculeLanguage model
How access works
Access begins with an AWS account and a payment method; the remaining friction is IAM rather than sign-up. Foundation models are subscribed through AWS Marketplace, so the calling role needs Marketplace subscribe and view permissions, and invoking a third-party model accepts that model's licence. GovCloud is enabled separately in both the GovCloud and the linked commercial account. Region selection is the customer's, making data residency a configuration decision rather than a vendor constraint.
- Self-serve
- Free tier
- Public price list
- API access
- SSO / SAML
- SOC 2 Type II
- ISO 27001
- HIPAA eligible
- GDPR DPA
- Region selectable
What it costs
Rates are published per model and per region, and the billing shape depends on the service. Hosted foundation models bill per million input and output tokens, with batch inference discounted for work that tolerates delay and provisioned throughput sold as hourly commitments. Managed training and hosting bill by instance type and duration, with a serverless option metered per millisecond. Genomics workflows meter separately again, per task-second and per gigabase-month of storage.
Last verified
