Qubrid AI · Pricing Plans

Qubrid Ai Plans Pricing

Qubrid AI sells GPU-accelerated AI infrastructure on a pay-as-you-go basis with three product surfaces - serverless OpenAI-compatible inference billed per token, GPU virtual machines rented on on-demand or 1-week / 1-month / 3-month / 1-year terms, and bare-metal NVIDIA HGX servers sold on annual contracts. New users receive a $1.00 free API credit on their first $5 deposit. Per-model token rates and per-GPU hourly VM rates are exposed inside the platform console; bare-metal annual prices are public.

4 Plans API Commons Plans
View Source
Artificial IntelligenceCloud ComputingGPUInferenceLarge Language Models

Plans

Pay-As-You-Go pay-as-you-go

OpenAI-compatible serverless inference billed per token. New accounts receive $1.00 free credit on their first $5 deposit. Per-model rates are listed inside the platform console.

Token Inference (token · usage) see https://platform.qubrid.com/playground USD
First-Deposit Credit (credit · usage) $1.00 free on first $5 deposit USD
GPU Virtual Machines usage-based

NVIDIA GPU virtual machines available on On-Demand, 1 Week, 1 Month, 3 Month, and 1 Year rental terms. Per-GPU hourly rates are listed in the Qubrid platform; longer commitments receive volume discounts.

GPU VM Rental (gpu-hour · usage) see https://www.qubrid.com/pricing USD
Bare Metal Servers subscription

Dedicated NVIDIA HGX 8-GPU bare-metal servers sold on annual contracts.

NVIDIA B300 8-GPU (Annual) (server · year) $380,000 USD
NVIDIA B200 8-GPU (Annual) (server · year) $297,840 USD
NVIDIA H200 8-GPU (Annual) (server · year) $210,240 USD
NVIDIA H100 8-GPU (Annual) (server · year) $133,414.80 USD
NVIDIA A100 8-GPU (Annual) (server · year) $89,500 USD
Enterprise enterprise

Custom-quoted engagements for larger deployments, custom fine-tuning, and enterprise solutions; arranged via the Qubrid sales team.

Custom Deployment (varies · usage) contact sales USD

Sources