tools
Baseten
Baseten provides managed infrastructure to deploy, serve, train, and distribute open-source, custom, and fine-tuned AI models.
Baseten lets teams deploy custom and open-source models, use pre-optimized Model APIs, run training jobs, and scale inference on managed GPU infrastructure. It supports OpenAI-compatible APIs, dedicated deployments, autoscaling, observability, and self-hosted deployments for enterprise customers.
Teams pay for compute or Model API usage. The Basic plan has no monthly platform fee but uses pay-as-you-go pricing; Pro and Enterprise pricing requires a quote. Baseten does not provide a general-purpose end-user chatbot.
Features
- Deploy open-source, custom, and fine-tuned models on dedicated infrastructure
- Use pre-optimized Model APIs with OpenAI-compatible endpoints
- Scale deployments with autoscaling and configurable GPU resources
- Run multi-node training jobs with checkpoint syncing
- Train with the Loops SDK and deploy checkpoints to inference
- Monitor deployments with logs, metrics, and observability APIs
- Use server-side hosted tools such as web search
- Package and serve models with the open-source Truss framework
Use cases
- Deploy custom models as production inference endpoints
- Integrate open-source Model APIs into an application
- Train and deploy models using managed GPU infrastructure
- Scale high-volume inference workloads across GPU instances
- Monitor model latency, logs, metrics, and usage
- Distribute models to customers through Baseten for Model Labs
Pros
Cons
Latest updates
- joey-1 decision model available on Baseten
joey-1 is now available through Baseten Model APIs, and its weights are open source on Hugging Face.
- elephant-1 and mammoth-1 pricing
elephant-1 and mammoth-1 are moving to paid usage on Baseten Model APIs starting October 27, 2026, at 11:59 p.m. PT.
- Per-replica metrics
You can now compare CPU, GPU, and memory usage across individual replicas in every workspace to spot outliers hidden by aggregate graphs.
- Revocation of unused API keys on 10/24
Baseten will automatically revoke API keys that haven’t authenticated a request in 30 days.
- Upcoming change to API authentication error responses
Baseten APIs will return 401 Unauthorized instead of 403 Forbidden when a request has missing, invalid, or expired credentials.
Capabilities
- Choice of models — “You can deploy open source and custom models on Baseten.” source
- Self-hosted — “We offer single-tenant and self-hosted deployments for extra security.” source
- API — “Model APIs” source
- Runs models for you — “Serve open-source, custom, and fine-tuned AI models on infra purpose-built for high-performance inference at massive scale.” source
Get it
Security
- SOC 2 Type II — “Baseten Labs, Inc. SOC 2 Type 2 Report 5.31.26.pdf” source
- SOC 2 — “Baseten Labs, Inc. SOC 2 Type 2 Report 5.31.26.pdf” source
- ISO 27001 — “Baseten Labs, Inc. ISO 27001 Certificate.pdf” source
- GDPR — “SOC 2 ISO 27001:2022 HIPAA CCPA GDPR PCI DSS - SAQ D” source
- HIPAA — “Baseten Labs, Inc. HIPAA Report 5.31.26.pdf” source
Pricing
- Prices checked
- 2026-11-05
- Read
- by web search
Pay as you go
- $0.035 per 1M input tokens (elephant-1)
- $0.10 per 1M input tokens (mammoth-1)
- $0.01052 per minute (T4)
- $0.0655 CPU/hr (Sandboxes compute)
- $0.0211 GB RAM/hr (Sandboxes memory)
- $0.20 GB per month (Sandboxes snapshots)
- Model APIs token pricing (input, cached input, output)
- Dedicated Inference compute per minute
- Training compute per minute
- Sandboxes runtime pricing
- $0 per month, pay as you go
- Fast cold starts, SOC 2 Type II and HIPAA compliant
Production scale
Price on request
- Volume discounts
- Reserved/priority compute
- Everything in Pay as you go plus custom terms