RanoAIAI Inference, Powered by NPU
Fast, efficient AI inference on Furiosa NPU hardware.
6 models · up to 120B params · 60% cheaper than GPU cloud.
# OpenAI-compatible — drop-in replacement
from openai import OpenAI
client = OpenAI(
base_url="https://api.ranoai.com/v1",
api_key="rno-xxxxxxxxxxxxxxxx"
)
# gemma-4-31b · gemma-4-26b · gemma-4-12b
# qwen-3.5-35b · qwen-3.5-27b
response = client.chat.completions.create(
model="gemma-4-31b",
messages=[{"role":"user",
"content":"Hello!"}]
)
# $0.25 / 1M input tokens
Next-Gen NPU
Built for Inference
NPU chips paired with Google and Alibaba's best open models — fast, affordable inference that GPU clouds can't match on cost.
Korea's second-generation NPU built for large-scale AI inference. RNGD delivers exceptional throughput per watt with a dataflow architecture purpose-built for transformer workloads.
GPT-OSS-120B, Gemma 4 (31B · 26B · 12B), and Qwen 3.5 (35B · 27B) — six state-of-the-art models with multimodal and long-context capabilities, all served on NPUs.
RanoAI NPU vs RTX 5090 & RTX PRO 6000
Gemma 4 27B inference · Furiosa RNGD (HBM3) vs NVIDIA Blackwell (GDDR7)
↓ lower is better for Power Draw and Token Cost · † int4 quantization required (fp16 exceeds 32 GB VRAM) · Source: Cloudrift, NVIDIA specs
From API Key to Response
in 4 Simple Steps
Drop-in replacement for OpenAI. Point to our endpoint and let NPUs handle the rest.
🔒 All connections are TLS-encrypted. Inference runs entirely on NPU infrastructure.
Get Your API Key
Sign up and instantly receive your API key from the dashboard. No waitlist, no approval process — start making requests within minutes.
Call the API
Use any OpenAI-compatible SDK. Point base_url to api.ranoai.com and swap in your key. Zero code changes needed if you're already on OpenAI.
NPU Runs Inference
Your request hits our NPU cluster running Furiosa RNGD chips. Inference completes with low latency at domestic data centers.
Monitor Usage
Track token consumption, latency, and cost in real time from your dashboard. Set spending limits and alerts to stay in full control.
Pay Only for
What You Use.
No subscriptions. No commitments. NPU efficiency means lower cost per token — no GPU markups.
| Model | Params | Context | Input / 1M tokens | Output / 1M tokens |
|---|---|---|---|---|
GPT-OSSgpt-oss-120b | 120B | 128K | $0.60 | $1.80 |
Gemma 4gemma-4-31b | 31B | 128K | $0.25 | $0.75 |
gemma-4-26b | 26B | 128K | $0.20 | $0.60 |
gemma-4-12b | 12B | 128K | $0.08 | $0.25 |
Qwen 3.5qwen-3.5-35b | 35B | 128K | $0.28 | $0.85 |
qwen-3.5-27b | 27B | 128K | $0.18 | $0.55 |
Need higher rate limits or dedicated capacity?