What nvidia models are in the catalog?
5 models from nvidia are listed with per-model pricing and endpoint data.
o10 State of Inference Spend 2026 found up to 638× compliant price spread across venues for identical workloads.
o10Last updated 2026-06-14b
nvidia models
5 models from nvidia in the unified inference model catalog.
Browse nvidia models with gateway pricing, provider endpoints, and eval-gated routing guidance.
Dashboards observe.
o10 enforces.
o10 enforces.
Every page in this index follows the same structure as the home site — answer-first, passage blocks, operational steps, and expanded FAQs. o10 State of Inference Spend 2026 found up to 638× compliant price spread across venues for identical workloads.
Start hereQuick overview
How to use this index
nvidia5 models
Nemotron 3 Nano 30B A3B
text · nvidia/nemotron-3-nano-30b-a3b
NVIDIA Nemotron 3 Super 120B A12B
text · nvidia/nemotron-3-super-120b-a12b
Nemotron 3 Ultra
text · nvidia/nemotron-3-ultra-550b-a55b
Nvidia Nemotron Nano 12B V2 VL
text · nvidia/nemotron-nano-12b-v2-vl
Nvidia Nemotron Nano 9B V2
text · nvidia/nemotron-nano-9b-v2
FAQFrequently asked questions
Common questions
How many nvidia models?
5 models in the o10 gateway catalog snapshot.
Where is pricing from?
Gateway catalog snapshot — verify against your gateway provider's published pricing.
How does o10 route nvidia models?
o10 State of Inference Spend 2026 found up to 638× compliant price spread across venues for identical workloads. o10 selects the cheapest nvidia tier that clears your per-use-case eval floor — starting in shadow mode.
o10Set the envelope. o10 holds it.
See what you're overpaying.
Paste a week of traffic. Get the number that books the audit.
See what you're overpaying →