Triton Inference Server · Pricing Plans
Triton Plans Pricing
NVIDIA Triton Inference Server is open-source software (BSD-3-Clause) that customers self-host on their own CPUs / GPUs. There is no per-API call price from NVIDIA for the server itself; commercial entitlements (support, indemnification, cloud images) ship via the separate NVIDIA AI Enterprise subscription.
2 Plans
API Commons Plans
View Source
AIInferenceOpen SourceModel Serving
Plans
Open Source
freemium
Self-hosted, freely redistributable Triton Inference Server. Cost to the user is compute (CPU / GPU) and operational overhead, not a license fee.
Software License (month · month)
0.00 USD
- BSD-3-Clause License
- Self-Hosted
- Community Support (GitHub Discussions)
- KServe V2 Inference Protocol
NVIDIA AI Enterprise (Optional Support)
enterprise
Optional commercial support, security patching, and indemnification for Triton delivered through the NVIDIA AI Enterprise subscription. Pricing is tied to the AI Enterprise product, not Triton calls.
Support Subscription (month · month)
see NVIDIA AI Enterprise pricing USD
- Enterprise Support
- Security Patching
- Indemnification