OpenAI compatible LLM router

The AI router tuned to save you money.

One API, one key and one prepaid balance for Anthropic, OpenAI, Google and 12 more makers. Send model auto and Zabio picks the cheapest model that can do the job, then shows you what it cost and what you saved.

Start free, no card See models and prices

Top-up matchWe match your top-ups, up to 250 USD, once you spend 250 USD in your first 30 days. How the match works
Referrals25 USD in credit for every developer you bring who adds 100 USD. How referrals work

Zabio/Sample team ▾All projects ▾maya@sample.team

Dashboard
Usage
Playground
API keys
Projects
Members
Models
Billing

Dashboard Sample team, this month

Spend$241.33What Zabio billed this month

Cost on baseline models$1,500.00The same calls on Claude Opus 5

Saved by routing$1,258.6783.9% less than the baseline

Calls100,00099,960 succeeded

Daily spend by model last 30 days

Claude models
Gemini models
Qwen models
OpenAI models

Cost breakdown

Routing saved you $1,258.67 (83.9%)

Provider price: $206.03
Service fee: $10.30
Picking the model for auto: $25.00
Billed: $241.33
On Claude Opus 5: $1,500.00

Watch the tourThe panel in about a minute, with sound

The panel in about a minute: a sample team's month in the dashboard, then organizations, projects, the team, usage, logs and the playground. Sample data: the calculator's typical 100,000 requests, sent to auto instead of Claude Opus 5.

Models from

Qwen
OpenAI
DeepSeek
Anthropic
Google
Mistral AI
Z.ai
deepinfra
Meta
Moonshot AI

and 5 more: every model and price

How it works

One API. The right model. Every cost on the record.

Change two lines
Point base_url at https://api.zabio.ai/v1 and use your Zabio key. The OpenAI SDKs for Python and Node work as they are, and so do tools built on them, such as LangChain and the Vercel AI SDK. Streaming, tools, images, JSON schema, embeddings and the Responses API all work.
```
from openai import OpenAI client = OpenAI(    base_url="https://api.zabio.ai/v1",    api_key="insert your zabio.ai key here",) reply = client.chat.completions.create(    model="auto",    messages=[{"role": "user", "content": "Say hello."}],)
```
Read the quickstart
Ask for auto, or name a model
Send model auto and Zabio picks the model for each request. Name a line, auto/best, auto/balanced or auto/cheap, to set the trade-off yourself, or name any model to skip routing.
model: autoSummarize this ticket in one sentence.
classifiersummarization, easy
- bestclaude-opus-5-5for the hardest jobs
- balancedgemini-3.8-flashfor medium jobs
- cheapqwen3.5-9banswered
How routing decides

See the cost and the saving

Every response carries a zabio field with the model that answered, why it was chosen, what the call cost and what it would have cost on your usual model. The dashboard adds it all up by model, key and member.

Time	Model	Job	Cost	Took
5:50 PM	`auto`claude-opus-5-5	Code generation	$0.0223	5.7 s
5:49 PM	`auto`qwen3.5-9b	Conversation	$0.0003	0.8 s
5:49 PM	`auto`gemini-3.8-flash	Classification	$0.0017	2.0 s
5:49 PM	`auto`gpt-6.1-sol	Data extraction	$0.0065	2.9 s

What every response carries

Budgets for keys, quotas for people
Give each API key a spend limit that resets by day, week or month, and each teammate a monthly quota. Once a key reaches its limit, its calls are refused until the limit resets, so a runaway loop costs what you allowed and no more.
- web-app132.73 of 300 USD
- support-bot72.40 of 150 USD
- nightly-batch36.20 of 80 USD
- Maya, no limit
- Luis, 150 USD a month
- Ana, 80 USD a month
Teams, keys and limits

Pricing in one line

Every cent routing saves is yours.

You pay the provider list price plus a (see the pricing page) service fee, and (see the pricing page)(see the pricing page) to (see the pricing page) per request when Zabio picks the model. We keep none of the saving.

Work out your month Full pricing

What a sample month costs, day by day

100,000 requests, 1,000 tokens in and 400 out each, a mix of easy, medium and hard jobs.

claude-opus-5 for everything1,500.00 USD
Through auto(see the pricing page)
What you keep, (see the pricing page) of the bill(see the pricing page)

Where the bill through auto goes

To the providers 205.92 USD
Service fee (see the pricing page)
Picking the model (see the pricing page)

Built for teams

Run every product and every person on one balance.

Organizations, projects and people

Keep each product in its own organization, group its keys into projects, and invite teammates with a role and a monthly quota. Usage splits the bill by model, key and member.

Sample teamOwner
Sample labsMember
Sample studioOwner
New organization

Member	Role	Monthly quota	Spent this month
Maya Chen (you)maya@sample.team	Owner	No limit	$120.67
Luis Ortegaluis@sample.team	Admin	$150.00	$77.23
Ana Ruizana@sample.team	Member	$80.00	$43.44

ProjectsWeb appSupport botData pipeline

A playground that shows its work

Send a prompt to auto and see which model answered, the job it saw and what the call cost. Then copy the same request as curl, Python, Node or PHP.

Try the playground

Playgroundmodel: auto

Draft a reply to this ticket: I was charged twice for October, but I only have one subscription. Please refund the extra charge.

Hi, thanks for flagging this. You were charged twice for October, so we have refunded the extra charge, and it should reach your card in 5 to 10 business days. Sorry for the trouble.

Model served gemini-3.8-flash
Job writing, medium
Cost $0.0012

Request as codecurlPythonNodePHP

What Zabio does

Routing saves the money. The rest keeps it under control.

Smart Splitting
One request, several models, a smaller bill, decided for you on every call. When we can save, we save.
Smart routing
The right model and reasoning setting for every request, picked from evidence, with the reason on every response.
Jev for classification
Tagging, routing, moderation and sentiment for a fraction of the price of a chat model, named or picked by auto.
Caching
Your history cached for you, the 1 hour cache when turns are slow, and never a cut of your cache discount.
Flex
Half price when you can wait, with one deadline field.
Batching
Thousands of requests in one file at about half price, with auto routing on every line.

All features

Two ways to earn Zabio credit

We match your top-ups, up to 250 USD. Spend 250 USD in the 30 days after you sign up, and from then on we match every top-up dollar for dollar until we have added 250 USD. Add 20, we add 20. Add 50, we add 50. How the match works

25 USD for every developer you bring. Share your link. When someone who signs up through it adds 100 USD, you get 25 USD in credit once their top-ups have cleared 30 days. Every referral counts. How referrals work

Offer credit has no cash value. A top-up that earns a match can't be refunded, and a disputed one loses its match. Read the terms

How Zabio compares

Pick OpenRouter for the widest catalog. Pick LiteLLM if the gateway must run inside your own infrastructure. Pick Portkey for deep observability and guardrails in a large company. Pick Zabio when you want the bill to go down and to see why each model was chosen.

Zabio vs OpenRouter
The widest model marketplace against a router that picks the model and keeps none of the saving. Where each one fits, and how to switch.
Zabio vs LiteLLM
An open-source proxy you run yourself against a hosted router with one balance for eight vendors.
Zabio vs Portkey
An AI gateway built for observability and governance against a router built to lower the bill.

The full comparison

Common questions

How is Zabio different from other AI routers?

It picks the model for you, to get you more right answers for less. Send auto and Zabio reads each request, picks the model and reasoning setting that fit the job, and shows why on every response, with what the call cost and what it saved. Pricing stays simple: the provider's list price, a small service fee on each call and a small fee on each auto request. Zabio takes no cut of what routing saves, so every cent of the saving is yours.

Is Zabio free?

Yes. Every account starts on the Free plan: 100 requests a day, up to 16,000 tokens in and 2,000 out per request, on six models from OpenAI, Google, Mistral and Together AI, with auto choosing among them. No card, no expiry. Every other model stays visible, one top-up away, and a daily fair-use limit keeps Free sustainable. Your first top-up removes every limit. Calls on your own vendor keys pay no service fee.

What is Smart Splitting?

Under auto, a request that carries a long document and asks about only part of it can be split: a cheap model copies out the lines the question needs, word for word, Jev checks that the extract holds everything the answer needs, and the model auto picked answers from that extract instead of the whole document. If the check fails, Zabio sends the full document and pays for the attempt itself, so a split never costs you more. In our tests it cut the cost of long-document questions by about a third on average, and by more than half on contracts, with every fact kept. It never applies to a model you name, every step shows in zabio.split, and zabio.split set to off turns it off for a request.

Can I switch models without losing the conversation?

Yes. With the Responses API, Zabio stores the conversation and sends the full history to whichever model answers the next turn, on any vendor. When the history outgrows that model's context window, Zabio summarizes the oldest part and continues. With chat completions your app sends the history on every call, so any model gets the full context.

Can I control what each person and each key spends?

Yes. Owners and admins set a monthly quota per member, and every API key can have its own monthly spend cap, request and token limits per minute, and a list of the models it may use. Calls stop with 429 quota_exceeded when a limit is reached.

All questions

Build your prototype free

Change two lines in your code, send model auto, and see what every call cost and what you saved.

Start free, no card Read the quickstart

Launch offer: spend 250 USD in the 30 days after you sign up and we match your top-ups after that, up to 250 USD. How it works