Workstation · FP8
$2.4/hr $0.0007/sec · $1,752/mo at 24/7
g4-standard-48 — try any app from $2.40/hr (park/wake receipted D27).
VRAM pool
1× NVIDIA RTX PRO 6000 Blackwell 96GB
Compute die
NVIDIA RTX PRO 6000 Blackwell 96GB — FP8 tensor cores.
GDDR7 memory
96 GB of GDDR7 across the GPU.
PCIe Gen5 interconnect
PCIe Gen5 host links.
Per-second metering
$2.4/hr → 0.0007/sec. Pause the deployment and the meter hits $0.
Scale-to-zero
Autoscale + region pins (GLOBAL / US / EU / APAC) on the control plane.
Memory budget
96 GB of VRAM on this preset. A 70B-class checkpoint fits (rule of thumb: about one byte per parameter at FP8), or several 13B workers side by side.
Typical loads
single-model 70B serving, team-shared dev capacity. Fine on: torch, JAX, raw CUDA, axolotl, vLLM, sglang, ComfyUI.
Billing math
$2.4/hr is $0.0007/sec while running and $0 while parked; park/wake freezes and restores state instead of paying for idle.
Best fit
96 GB of GDDR7 on one RTX PRO 6000 Blackwell card over PCIe Gen5, $2.40/hr. It is the first rung with enough VRAM for a 70B checkpoint on a single device: at FP8 and one byte per parameter that is ~70 GB of weights and ~26 GB left for KV cache; a 30B model leaves ~66 GB (rule of thumb).
Trade-off
g4-2x is $4.80/hr: a second 96 GB card and 192 GB in total, still $2.40/hr per GPU — the same per-GPU price as here. Each card has its own PCIe Gen5 link, so the step up adds a second independent 96 GB budget, not more VRAM per device.
$2.4/hr on runaii.cloud, metered per second: $0.0007/sec while the VM runs and $0 while parked. A month of 24/7 uptime would be about $1,752, and park/wake means busy-fraction pricing in practice.
Yes. Per-second metering with park/wake: idle freezes the VM state into a signed artifact and stops the meter the same second; the next request wakes it byte-exact
96 GB of VRAM. A 70B-class checkpoint fits (rule of thumb: about one byte per parameter at FP8), or several 13B workers side by side.
A 70B checkpoint at about one byte per parameter leaves ~26 GB of the 96 GB for KV cache and activations; a 30B model leaves ~66 GB (rule of thumb).
No — one card. Multi-GPU RTX PRO 6000 starts at g4-2x (2× 96 GB, $4.80/hr) and runs to g4-8x (8× 96 GB, $19.20/hr).
An L4 is 24 GB of GDDR6 on PCIe Gen4 at $0.87/hr; the RTX PRO 6000 is 96 GB of GDDR7 on PCIe Gen5 at $2.40/hr — four times the VRAM per card and 2.8× the hourly.
RTX PRO 6000 slices
from $0.30/hr · 12-48GB
Blackwell slices for light jobs
1× NVIDIA T4 16GB
$0.44/hr · FP16
Starter
1× NVIDIA L4 24GB
$0.87/hr · FP8
Efficient
2× NVIDIA L4 24GB
$1.8/hr · FP8
Efficient ×2
4× NVIDIA L4 24GB
$3.6/hr · FP8
Efficient ×4
8× NVIDIA L4 24GB
$7.2/hr · FP8
Efficient ×8
2× NVIDIA RTX PRO 6000 Blackwell 96GB
$4.8/hr · FP8
Workstation ×2
4× NVIDIA RTX PRO 6000 Blackwell 96GB
$9.6/hr · FP8
Workstation ×4
8× NVIDIA RTX PRO 6000 Blackwell 96GB
$19.2/hr · FP8
Workstation ×8
8× NVIDIA B300 288GB
$96/hr · FP8
Full Precision
8× NVIDIA B200 180GB
$80/hr · NVFP4
Throughput
4× NVIDIA B300 288GB
$48/hr · NVFP4
Minimal
4× NVIDIA B200 180GB
$40/hr · NVFP4
Efficient
16× NVIDIA GB300
$184/hr · FP8
Frontier Max