Pricing
Get access to the latest open-source models and the best GPUs on the market
Model Platform
Serverless Inference
Price per 1M tokens
| Model | Input | Output | Cache | |
|---|---|---|---|---|
| $0.26 | $0.40 | $0.052 | Try Model API | |
| $0.27 | $1.08 | $0.07 | Try Model API | |
| $0.27 | $1.08 | $0.03 | Try Model API | |
| $0.50 | $2.80 | $0.10 | Try Model API | |
| $0.60 | $3.60 | $0.55 | Try Model API | |
| $0.08 | $0.24 | $0.04 | Try Model API | |
| $0.90 | $3.10 | $0.20 | Try Model API | |
| $0.43 | $1.75 | $0.11 | Try Model API |
GPU cloud
GPUs
NVIDIA GB200 NVL72| Instance Type | vCPUs | RAM | NVMe Storage | Public IP | Price/Hour |
|---|---|---|---|---|---|
| NVIDIA GB200 NVL72 | 144 | 960 GB | 28 TB | 1 | $9.00 |
NVIDIA HGX B200| Instance Type | vCPUs | RAM | NVMe Storage | Public IP | Price/Hour |
|---|---|---|---|---|---|
| 2x NVIDIA SXM5 B200 | 60 | 400 GB | 2 TB | 1 | $4.50 |
| 4x NVIDIA SXM5 B200 | 120 | 800 GB | 4 TB | 1 | $4.50 |
| 8x NVIDIA SXM5 B200 | 240 | 1.5 TB | 8 TB | 1 | $4.50 |
NVIDIA HGX H200| Instance Type | vCPUs | RAM | NVMe Storage | Public IP | Price/Hour |
|---|---|---|---|---|---|
| 2x NVIDIA SXM5 H200 | 40 | 400 GB | 2 TB | 1 | $3.00 |
| 4x NVIDIA SXM5 H200 | 80 | 800 GB | 4 TB | 1 | $3.00 |
| 8x NVIDIA SXM5 H200 | 160 | 1.5 TB | 8 TB | 1 | $3.00 |
NVIDIA HGX H100| Instance Type | vCPUs | RAM | NVMe Storage | Public IP | Price/Hour |
|---|---|---|---|---|---|
| 2x NVIDIA SXM5 H100 | 40 | 400 GB | 2 TB | 1 | $2.25 |
| 4x NVIDIA SXM5 H100 | 80 | 800 GB | 4 TB | 1 | $2.25 |
| 8x NVIDIA SXM5 H100 | 160 | 1.5 TB | 6 TB | 1 | $2.25 |
Storage & Networking
Storage| Product | Price |
|---|---|
| Network Shared Storage | $0.07 per GB per month |
| Object Storage | $0.04 per GB per month |
Networking| Product | Price |
|---|---|
| Additional public IP address | $4.00 per IP per month |
