NeoAI API

Every model.
One API.

The leading models from every lab behind one OpenAI-compatible endpoint. Call any model by its id, or call Neo, the layer that unifies them all. Pay only for what you use.

Get an API key Read the docs

90+ models20+ labsCredits from $10OpenAI-compatible

neoolympus.ai/v1

$ curl https://neoolympus.ai/v1/chat/completions \
    -H "Authorization: Bearer $NEO_API_KEY" \
    -d '{"model": "neo-terra", "messages": [{"role": "user", "content": "Summarise this contract in one line"}]}'

{
  "object": "chat.completion",
  "model": "neo-terra",
  "choices": [{"message": {"role": "assistant", "content": "A 24-month supply deal, renewable, with 60 days' notice."}}],
  "usage": {"prompt_tokens": 1840, "completion_tokens": 21, "total_tokens": 1861},
  "neo": {"level": "terra", "cost_usd": 0.00177}
}

Every model, one API.

One key, one bill and one format for the whole frontier. Switch models by changing one word.

model: "neo-terra"Neo, one layer for every model

Neo unifies the whole frontier behind one name, in three levels with one fixed price each: Neo Luna, Neo Terra and Neo Sol.

model: "<id>"Or pick any model

Call the latest models from OpenAI, Anthropic, Google, xAI, DeepSeek, Mistral, Qwen and more by their id.

toolsTool calling

Every model is tested for tool calling and the result is in the model list. Neo sends tool calls only to models that pass.

stream: trueStreaming

Answers arrive token by token, as server-sent events in the OpenAI format.

neo_zdr: trueZero data retention

Only providers that keep nothing of your request answer it.

fallbackAutomatic fallback

If a model is down or slow, the next best one answers, so your product keeps working.

images · audio · videosPictures, video and voice

Neo makes pictures, video and voice and transcribes audio, in the chat or through their own endpoints. Send pictures and audio in too.

neo.cost_usdCost in every answer

Each response says what it cost. The console shows usage by key, by model and by your own customers.

Start building in minutes.

Already using the OpenAI SDK? Change the base URL and the key. Nothing else.

1Create your account
Open the console, sign in with your NeoAI account and accept the API terms.
2Add credits
From $10, by card. Credits never expire.
3Make a key
It is shown once: keep it on your server and send your first request.

import os
from openai import OpenAI

client = OpenAI(
    base_url="https://neoolympus.ai/v1",
    api_key=os.environ["NEO_API_KEY"],
)

r = client.chat.completions.create(
    model="neo-terra",  # neo-luna, neo-sol or any model id from /v1/models
    messages=[{"role": "user", "content": "Explain quantum computing in two sentences."}],
)
print(r.choices[0].message.content)

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://neoolympus.ai/v1",
  apiKey: process.env.NEO_API_KEY,
});

const r = await client.chat.completions.create({
  model: "neo-terra", // neo-luna, neo-sol or any model id from /v1/models
  messages: [{ role: "user", content: "Explain quantum computing in two sentences." }],
});
console.log(r.choices[0].message.content);

curl https://neoolympus.ai/v1/chat/completions \
  -H "Authorization: Bearer $NEO_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "neo-terra",
    "messages": [{"role": "user", "content": "Explain quantum computing in two sentences."}]
  }'

Neo, at three levels.

Three Neo models that do everything: text, pictures, video, voice and transcription. Each has one fixed price per unit, whichever model answers.

Neo Luna

neo-luna

Fast and cheap: short answers, extraction, classification.

Input / 1M tokens: —
Output / 1M tokens: —
Image: —
Video / second: —
Voice / 1M characters: —
Transcription / minute: —

Default

Neo Terra

neo-terra

The default: writing, analysis and everyday code.

Input / 1M tokens: —
Output / 1M tokens: —
Image: —
Video / second: —
Voice / 1M characters: —
Transcription / minute: —

Neo Sol

neo-sol

The strongest models: hard reasoning, long documents, complex code.

Input / 1M tokens: —
Output / 1M tokens: —
Image: —
Video / second: —
Voice / 1M characters: —
Transcription / minute: —

US dollars. Video up to 720p. Set a maximum cost for any single request with neo_max_cost_usd.

Models and pricing.

Every model at its price, per million tokens, per picture, per second or per minute: the provider's price plus NeoAI's fee. Prices follow the providers' and update on their own.

Model	Context	Input / 1M	Output / 1M	Features
Loading models…

US dollars. Tax is added at checkout where it applies. ZDR: zero data retention available.

Choose how to start.

Build on your own with prepaid credits, or talk to us about a monthly invoice and higher limits.

Pay as you go

Everything you need to build and ship.

$10first top-up

Every model, and Neo at three levels
Prepaid credits that never expire
Usage by key, by model and by your customers
A monthly spend cap you set
120 requests a minute per key

Get an API key

Enterprise

For products built on NeoAI at scale.

Custom

Monthly invoice
Higher rate and concurrency limits
Test keys with their own budget
Zero data retention for every request
Only the models you approve
A direct line to the team

Questions

Still have a question? Read the API reference

Is it compatible with the OpenAI SDKs?+

Yes. The API speaks the OpenAI Chat Completions format, with streaming and tool calling. Point any OpenAI SDK at https://neoolympus.ai/v1 and use your NeoAI key.

What do the Neo models do?+

Neo is one layer over every model: send it text, code or images and choose the level, Neo Luna, Neo Terra or Neo Sol. If a model fails, another one answers. The response shows the level and the cost.

Can Neo make pictures, video and voice?+

Yes. Ask for them in the chat, or call /v1/images, /v1/audio and /v1/videos. Audio you send is transcribed, and with modalities audio the answer is also read aloud. Each Neo model has a fixed price per picture, per second of video, per million characters and per minute of audio.

How is the price calculated?+

Each request costs the tokens it uses at its price per million tokens, as listed above. Neo Luna, Neo Terra and Neo Sol have one fixed price each, whichever model answers. Every response includes its cost.

What happens when my credits run out?+

Requests are refused with the error insufficient_credits until you add more. Nothing is ever charged automatically.

Do credits expire?+

No. Unused credits stay in your account. Every top-up and its invoice are in the console.

Do you train on my data or keep my prompts?+

No. NeoAI does not train on your data and does not keep the content of API requests: it records only what billing needs (time, model, tokens and cost). With zero data retention on, only providers that keep nothing answer.

What are the limits?+

By default, 120 requests a minute per key and 4 requests at a time per account. Need more? Write to info@neoolympus.ai.

Is tax included?+

Prices exclude tax. VAT or sales tax is added at checkout where it applies, and businesses can add their tax ID.

Every model.One API.