H100
80GB vRAM
- HBM3 · 3.35 TB/s
- Flagship throughput
from$2.79/hr
< 5 min
From catalog to live endpoint
OpenAI
Drop-in compatible API
per-hour
Billed only for what you run
vLLM
Production inference runtime
Everything to ship inference
From model to production endpoint
A focused path from an open-source model to a live, billable API — without standing up GPUs, runtimes, or gateways yourself.
Drop-in compatible
Already speak OpenAI? Just change the base URL.
Drop-in compatible with the OpenAI SDK. Swap the base URL, keep your code.
curl https://api.nusapod.io/v1/chat/completions \
-H "Authorization: Bearer $NUSAPOD_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"qwen2.5-7b-instruct","messages":[{"role":"user","content":"Hello"}]}'Per-hour GPU rental
Rent the GPU you need
From flagship H100s to value-tier 4090s — pick the card that fits your model and your budget, and pay only for the hours you run.
A100
80GB vRAM
- HBM2e · 2.0 TB/s
- Proven workhorse
from$1.89/hr
L40S
48GB vRAM
- GDDR6 · 864 GB/s
- Cost-efficient inference
from$0.99/hr
RTX 4090
24GB vRAM
- GDDR6X · 1.0 TB/s
- Best price per token
from$0.59/hr
Indicative pricing — final rates vary by region.
How it works
From catalog to live endpoint in three steps
Model catalog
Curated, ready to deploy
Pick a model and it runs on vLLM behind an OpenAI-compatible endpoint in minutes.
deepseek-r1-671bdeepseek-v3-671bllama-3.1-405b-instructllama-4-maverick-400bqwen3-235b-a22bmixtral-8x22b-instructdbrx-instruct-132bcommand-r-plus-104bqwen2.5-72b-instructllama-3.3-70b-instructqwen2.5-coder-32bgemma-2-27b-itmistral-7b-instructllama-3.1-8b-instructqwen2.5-7b-instructphi-4+ bring your own