RunInfraby RightNow

Dashboard Sign in Get started

Backed by

Combinator

Open models, built for agents

Fast responses. Lower-cost cached context. One API key for the OpenAI and Anthropic SDKs.

Get an API key Browse the models

Lower cached input price (USD / 1M tokens)

Your session stays warm

Session-aware routing keeps repeat calls close to cached context.

Without routing, a repeat lands warm one time in four.

Without routing, a repeat lands warm one time in four.

One session, one replica.
Repeat calls prefer the replica holding their context.
No hint needed.
Provide a session ID, or let automatic affinity use the conversation when available.
Reported, billed as cached.
Cache hits appear in usage and are billed at the cached rate. Hits stay best effort.

What agents actually pay

Per 1M input tokens, at each model's measured cache hit rate.

Price per 1M input tokens, log scale

$0.01$0.02$0.05$0.10$0.20

GLM 5.3 Flash
List price $0.10Effective input $0.011
9.6x
cheaper, 99.6% cache hit
Nemotron 3.5 Lightning 30B
List price $0.05Effective input $0.012
4.2x
cheaper, 95.3% cache hit
Qwen3.8 27B
List price $0.10Effective input $0.013
7.7x
cheaper, 96.7% cache hit

Streams report cached tokens when usage is requested. Your usage page shows your own billed figures.

4 models

Capabilities

FiltersAll models

Capabilities

GLM 5.3 Flash

Model capabilities: Tool calling / JSON mode / Streaming / Image input

per 1M input tokens

per 1M cached input tokens

per 1M output tokens

cache hit, last 24 hours

output tok/s, model only

Effective input $0.011 per 1M at that hit rate Effective input rate $0.011 per 1M tokens, from the measured cache share on billed agent traffic in the last 24 hours.

Nemotron 3.5 Lightning 30B

Model capabilities: Tool calling / JSON mode / Streaming

per 1M input tokens

per 1M cached input tokens

per 1M output tokens

cache hit, last 24 hours

output tok/s, model only

Effective input $0.012 per 1M at that hit rate Effective input rate $0.012 per 1M tokens, from the measured cache share on billed agent traffic in the last 24 hours.

Qwen3.8 27B

Model capabilities: Tool calling / JSON mode / Streaming / Image input

per 1M input tokens

per 1M cached input tokens

per 1M output tokens

cache hit, last 24 hours

output tok/s, end to end

Effective input $0.013 per 1M at that hit rate Effective input rate $0.013 per 1M tokens, from the measured cache share on billed agent traffic in the last 24 hours.

Ornith 1.5 35B

Model capabilities: Tool calling / JSON mode / Streaming / Image input

per 1M input tokens

per 1M cached input tokens

per 1M output tokens

cache hit, not yet published

output tok/s, model only

Pick a model. Copy the call.

Works with the OpenAI and Anthropic SDKs.

Get your API key

ShellChat completion

CLIENT_REQUEST_ID="$(uuidgen)"

curl -X POST "https://api.runinfra.ai/v1/chat/completions" \
  -H "Authorization: Bearer $RUNINFRA_GATEWAY_KEY" \
  -H "X-Client-Request-Id: $CLIENT_REQUEST_ID" \
  -H "Content-Type: application/json" \
  -d '{"model":"nemotron-3-5-lightning-30b","messages":[{"role":"user","content":"Hello"}],"max_tokens":4096}'

import os
import uuid
from openai import OpenAI

api_key = os.environ.get("RUNINFRA_GATEWAY_KEY")
if not api_key: raise RuntimeError("Set RUNINFRA_GATEWAY_KEY first.")

client = OpenAI(
    api_key=api_key,
    base_url="https://api.runinfra.ai/v1",
)

response = client.chat.completions.create(
    model="nemotron-3-5-lightning-30b",
    messages=[{"role": "user", "content": "Hello"}],
    max_tokens=4096,
    extra_headers={"X-Client-Request-Id": str(uuid.uuid4())},
)
print(response.choices[0].message.content)

import { randomUUID } from "node:crypto";
import OpenAI from "openai";

const apiKey = process.env.RUNINFRA_GATEWAY_KEY;
if (!apiKey) throw new Error("Set RUNINFRA_GATEWAY_KEY first.");

const client = new OpenAI({
  apiKey,
  baseURL: "https://api.runinfra.ai/v1",
});

const response = await client.chat.completions.create(
  {
    model: "nemotron-3-5-lightning-30b",
    messages: [{ role: "user", content: "Hello" }],
    max_tokens: 4096,
  },
  { headers: { "X-Client-Request-Id": randomUUID() } },
);

console.log(response.choices[0]?.message.content ?? "");

// Server-side only: never expose RUNINFRA_GATEWAY_KEY in browser code.
import { createOpenAICompatible } from "@ai-sdk/openai-compatible";
import { generateText } from "ai";

const apiKey = process.env.RUNINFRA_GATEWAY_KEY;
if (!apiKey) throw new Error("Set RUNINFRA_GATEWAY_KEY first.");

const runinfra = createOpenAICompatible({
  name: "runinfra",
  baseURL: "https://api.runinfra.ai/v1",
  apiKey,
});

const { text } = await generateText({
  model: runinfra.chatModel("nemotron-3-5-lightning-30b"),
  maxOutputTokens: 4096,
  messages: [
    { role: "user", content: "Write a TypeScript retry helper." },
  ],
});

console.log(text);

import os
from anthropic import Anthropic

client = Anthropic(
    # The SDK appends /v1, so use the bare host.
    base_url="https://api.runinfra.ai",
    api_key=os.environ.get("RUNINFRA_GATEWAY_KEY"),
)

message = client.messages.create(
    model="nemotron-3-5-lightning-30b",
    max_tokens=4096,
    messages=[{"role": "user", "content": "Hello"}],
)
print(message.content[0].text)

Zero data retention by default.

If you need a specific model in Model APIs, describe what you need