CPU instances
1st instance of either 2x2, 2x4 and 2x8 is free for 1 month
Instance Name | vCPU | RAM | Hourly | Spot * |
---|---|---|---|---|
2x2 | 2 | 2 | $0.02 | $0.01 |
2x4 | 2 | 4 | $0.05 | $0.02 |
2x8 | 2 | 8 | $0.10 | $0.04 |
4x16 | 4 | 16 | $0.20 | $0.06 |
8x32 | 8 | 32 | $0.40 | $0.18 |
16x32 | 16 | 32 | $0.82 | $0.31 |
GPU instances
Every new Brev user gets 30 minutes of free GPU use
GPU Model | vCPU | RAM | vRAM | Hourly | Spot * |
---|---|---|---|---|---|
1 x AMD V520 | - | - | 8 | $0.45 | $0.14 |
1 x Nvidia T4 | - | - | 16 | $0.63 | $0.20 |
1 x Nvidia A10 | 30 | 200 | 24 | $0.72 | - |
1 x Nvidia K520 | - | - | 8 | $0.78 | $0.24 |
1 x Nvidia M60 | - | - | 8 | $0.90 | $0.30 |
1 x Nvidia K80 | - | - | 12 | $1.08 | $0.38 |
1 x Nvidia A10G | - | - | 24 | $1.21 | $0.37 |
1 x Nvidia A100 | 30 | 200 | 40 | $1.31 | - |
2 x AMD V520 | - | - | 16 | $2.08 | $0.63 |
2 x Nvidia A100 | 60 | 400 | 80 | $2.64 | - |
2 x Nvidia M60 | - | - | 16 | $2.74 | $0.83 |
2x Nvidia K520 | - | - | 16 | $3.12 | - |
1 x Nvidia V100 | - | - | 16 | $3.67 | $1.11 |
4 x AMD V520 | - | - | 32 | $4.16 | $1.25 |
4 x Nvidia T4 | - | - | 64 | $4.69 | $1.53 |
4 x Nvidia A100 | 120 | 800 | 160 | $5.28 | - |
4 x Nvidia M60 | - | - | 32 | $5.47 | $1.65 |
4 x Nvidia A10G | - | - | 96 | $8.81 | $2.05 |
8 x Nvidia K80 | - | - | 96 | $8.64 | $2.60 |
16 x Nvidia K80 | - | - | 192 | $17.28 | $5.19 |
8 x Nvidia A10G | - | - | 192 | $19.55 | $5.87 |
8 x Nvidia A100 | 124 | 1800 | 320 | $10.56 | - |
8 x Nvidia A100 | - | - | 320 | $39.33 | $11.80 |
- Estimated based on previous Spot prices
Storage
$0.0001314 per GB per hour for the lifetime of an environment
No need to always be on a GPU!
Every Brev environment can be scaled on the fly. No need to stick to a GPU instance if you don't need it. Behind the scenes, we do this by shifting your machine's disk over to a different instance type, briefly restarting it.
Scaling an environment
In the brev console, go to your environment settings and under machine settings, you'll see a slider to scale your instance. Or via the CLI run:
brev scale [ Workspace Name or ID... ] --cpu 2x8
(At the moment, this only works for non-SPOT instances. We are working on a fix for SPOT instances so stay tuned 🤙)