Minimal · NVFP4
$48/hr $0.0133/sec · $35,040/mo at 24/7
Smallest frontier-capable footprint.
VRAM pool
288GB ×4 = 1.1TB
Position
vs H200 4×: lower $/hr, more VRAM
Compute die
NVIDIA B300 288GB — NVFP4 tensor cores, tuned for smallest frontier-capable footprint.
HBM3e memory
1,152 GB of HBM3e across the 4 GPUs.
NVLink 5 interconnect
NVLink 5 across the 4 GPUs — one memory domain.
Per-second metering
$48/hr → 0.0133/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
1,152 GB of VRAM on this preset. Frontier MoE serving at full batch with KV-cache headroom (rule of thumb: about half a byte per parameter at NVFP4).
Typical loads
large MoE serving, high-QPS inference, batch evals at scale. Fine on: torch, JAX, raw CUDA, axolotl, vLLM, sglang, ComfyUI.
Billing math
$48/hr is $0.0133/sec while running and $0 while parked; park/wake freezes and restores state instead of paying for idle.
Best fit
1,152 GB across four B300 288 GB cards on NVLink 5 at $48/hr, $12 per GPU-hour — the catalog's smallest frontier-capable footprint. At NVFP4 a 1T-parameter checkpoint is ~500 GB of weights and leaves ~650 GB of the 1,152 GB pool; even at FP8 it is ~1 TB, most of the pool (rule of thumb).
Trade-off
b200-8x-fp4 is $80/hr: eight B200 180 GB cards, 1,440 GB total, $10 per GPU-hour. That is $32/hr more than this preset for twice the cards and 25% more VRAM, but 180 GB per card against 288 GB here — a job that needs 288 GB on one device stays on the B300 row.
$48/hr on runaii.cloud, metered per second: $0.0133/sec while the VM runs and $0 while parked. A month of 24/7 uptime would be about $35,040, 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
1,152 GB of VRAM. Frontier MoE serving at full batch with KV-cache headroom (rule of thumb: about half a byte per parameter at NVFP4).
1,152 GB — four 288 GB B300 HBM3e cards on NVLink 5 at $48/hr, $12 per GPU-hour.
Same four-card shape, $8/hr more ($48 vs $40) and 1,152 GB instead of 720 GB: 288 GB per card against 180 GB, both at NVFP4.
It is the lowest-priced B300 shape in the catalog at $48/hr for four cards; the catalog's note for the row is exactly that phrase.
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
1× NVIDIA RTX PRO 6000 Blackwell 96GB
$2.4/hr · FP8
Workstation
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 B200 180GB
$40/hr · NVFP4
Efficient
16× NVIDIA GB300
$184/hr · FP8
Frontier Max