Baseten
www.baseten.co- Category
- Developer Tools
- Rank
- No. 1284Tools index
Previous survey · No. 1257 ·
- Listed in
- #5 Choose compute and GPUs
- Pricing
- Paid
- Type
- TOOL
- Builder
- @baseten
- Date
About
Dedicated model deployments on GPUs. Deploy and autoscale models on dedicated GPUs with a managed serving stack. Billed per GPU-minute of active capacity.
What it does
According to its pricing page, Baseten lets teams deploy open-source or custom models on dedicated GPUs, call pre-optimized models through hosted APIs, and run training jobs, with tiers ranging from pay-as-you-go to enterprise plans that support self-hosting in a customer's own cloud.
Stated on the product site
- Compliance certifications
- The site states its platform is SOC 2 Type II and HIPAA compliant.
- Idle time billing
- The FAQ states customers are not charged while a model is idle, only while it is deploying, scaling, or serving predictions.
- Self-hosting option
- The Enterprise plan description lists self-hosted deployments in the customer's own cloud as an option.
- Model packaging framework
- Custom models can be packaged and served using Truss, which the site calls its open source standard for packaging and serving models.
Not stated on the site
- The page does not specify which cloud regions or providers are available under the Basic plan.
- Exact volume discount terms and enterprise pricing are not disclosed on the page and require contacting sales.
- The page does not describe SLA guarantees for the Basic or Pro tiers, only that Enterprise includes custom SLAs.
Written from the product site at baseten.co.
Intel on Baseten
Tags
BasetenGPU Cloud
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.