DECISIONS API
MODEL RELEASE / 3.4Meet Vansa-3.4. More context for your decisions.
APPLIED AI / DECISION MODELS

Intelligence,
with direction.

AI for the decisions inside your software. Give Vansa context and possible outcomes. Get back a choice, a score, or a probability you can act on.

Try Vansa for free Start with Google

VANSA-3.4 / NOW AVAILABLE

FIG. 01CONTEXT → DECISION
A three-dimensional dotted landscape rising toward one rust-colored decision peak.
DEFINED OUTCOME
From a space of possibilities,
a defined next step.CONTEXT
TO DECISION

CURRENT MODELVansa-3.4

Less manual workClassify and route requests automatically.
Low running costsPay for input. No output fees.
More controlChoose when to act or review.

VANSA VS GENERATIVE LLMS

Fast decisions.
Small costs.

Your software needs to route, classify, or prioritize. Vansa scores the answers you define directly, skipping the text-generation step.

BUILT FOR SPEED

No text to generate.

Get a decision without waiting for an LLM to write an answer token by token.

LOW COST, BY THE TOKEN

$0.034 USD / 1M input tokens

No generated output tokens to pay for. Prepaid credit from $1, or a monthly plan from $5.

Explore pricing

READY FOR YOUR SOFTWARE

Answers + probabilities.

Get a choice, a score, or yes/no. Use probabilities to act or send uncertain cases for review.

SAME INPUT. SAME QUESTION.

“Checkout is failing. Customers cannot pay.”

How urgent is it?Routine / Soon / High / Critical

GENERATIVE LLM

Generate an answer.

  1. 01
    Read the contextUnderstand the input and instructions.
  2. 02
    Generate the responseProduce answer tokens in sequence.
RETURNSText or structured JSON

A flexible choice for writing, coding, and open-ended tasks. Can also classify and return structured answers.

VANSA / DECISION MODEL

Score your options.

  1. 01
    Read the contextUnderstand the input and your question.
  2. 02
    Score the defined answersEvaluate your options without generating text.
RETURNSAnswer + probabilities

Purpose-built for the decisions inside your app: classification, prioritization, routing, and the next action.

HOW IT WORKS

One model.
Many everyday decisions.

Rank search results. Check AI answers. Organize documents. Choose what happens next in an app or a game. Explore the same three steps across different kinds of software.

CONTEXT → DECISION → ACTION

IN YOUR HELPDESK

Get each request to the right team.

  1. 01

    Send the context.

    A customer email, illustrated as a message in an envelope.

    “We were billed twice for March.”

    An excerpt from a customer’s email. Send the full message as context.

  2. 02

    Ask Vansa.

    BillingTechnicalSalesOther

    “Which team should handle this?”

    Define your teams as the possible answers. Vansa scores each option.

  3. 03

    Your app acts.

    A ticket arriving in the Billing folder, with a check mark.

    Billing99.2% probability

    Your helpdesk uses the answer to route the ticket to Billing.

Your context. Your choices. A decision your software can use.

Try this use case
View the full API response

One email can answer several questions. Switch between the team, urgency, and risk of the customer leaving to see each result.

CONTEXT → DECISIONSupport request
01 / INPUT STATEsupport_ticket.json
Duplicate charge on invoice #4411

“Hi, we were billed twice for March. Please refund the duplicate today or we will cancel our plan.”

vansa-3.4 / System One
02 / STRUCTURED ANSWERSJSON
billing99.2%probability
billing99.22%
technical0.52%
sales0.09%
other0.17%
Typed answers. Ready for your code.0 generated tokens
One state. Multiple questions. One API call.

OUTPUT FORMATS

Three ways
to find your answer.

Classify documents, score relevance, or check an AI answer. Get results in the shape your application needs.

choice

Pick the right option.

Define your categories. Get a selection and the probability of every option, with answers constrained to your set.

departmentbilling
Ticket routing · Intent detection
score

Put it on a scale.

Describe your ordered levels. Get an expected score and the distribution behind it, from low priority to critical.

urgency1.70 / 2
Urgency scoring · Sentiment analysis
noul

Ask a yes-or-no question.

Ask whether a statement holds. Get the probability of yes, so your application can act at the threshold you choose.

churn_risk85.68% yes
Email triage · Content checks

Example values: Vansa-3's real answers to the documented support-ticket request. Up to 32 questions per request; see API limits.

DEVELOPER INTERFACE

A little JSON.
A clear answer.

One endpoint for classification, scoring, and yes-or-no decisions. Use the tools you already work with.

  1. 01
    Send your state

    Text, a JSON object, or a conversation.

  2. 02
    Define your questions

    Set the instructions and allowed options.

  3. 03
    Use the answers

    Read typed results and probabilities.

Explore the API reference
curl https://api.vansa.org/v1/systemone \
  -H "Authorization: Bearer $VANSA_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "vansa-latest",
    "state": "Please refund my duplicate charge.",
    "questions": {
      "intent": {
        "type": "choice",
        "instructions": "What does the customer need?",
        "criteria": ["refund", "support", "other"]
      }
    }
  }'
POST /v1/systemoneBearer auth

Create an API key in the console. Keep it on your server.

FIELD NOTES

A closer look.

The model, the API, and the details that matter before you build.

01What is Vansa?
Vansa is a decision model API. You send a state (text or JSON) and typed questions (choice, score or yes/no), and it returns a calibrated probability for every option you defined, without generating text.
02What's new in Vansa-3.4?
Released 28 September 2026, Vansa-3.4 is the current model and supports a 16,384-token context per question. Earlier model names continue to work and are answered by Vansa-3.4. What changed
03How is it different from an LLM?
Chat models such as ChatGPT or Claude write text. Vansa scores the options you define without generating text, so every answer is one of your options, with a probability, and the same request gives the same answers.
04Who is it for?
Developers and teams building decisions into their software: search ranking, recommendations, document pipelines, AI answer checks, customer operations, agent routing, and interactive games. Send the context, define the possible outcomes, and use the returned scores in your application.
05How accurate is it?
Accuracy depends on your inputs and questions. Try representative examples in the playground, validate results on your own data, and use confidence scores to send uncertain cases for review.
06What does it cost?
Prepaid credit costs $0.034 USD per million input tokens, in any amount from $1 to $1,000, and purchased tokens carry over until used. Monthly plans: Starter, $5 for 150 million input tokens per month; Growth, $25 for 800 million; Scale, $99 for 3.5 billion. Unused monthly tokens expire at renewal. Output tokens are free and the public playground is free to try. New verified accounts can receive a one-time welcome bonus of 10,000,000 tokens. One per person; a bonus can be held for review, for example for duplicate, disposable or automated sign-ups. Pricing & usage calculator
07Is it compatible with Jev?
Yes. POST /v1/systemone accepts the same state and questions format as TypeSafe's Jev, so Jev client code works after changing the base URL, the key and the model name.
08What happens to my data?
Requests pass through Cloudflare to Vansa-owned infrastructure, where our model runs. We do not store full API request bodies or answers in the routine usage database. Operational records include request metadata, usage and client IP addresses; diagnostic logs may contain fragments of submitted content. Playground drafts are saved separately in your browser. We do not send API inputs to a third-party AI inference provider. See our Privacy Policy for details.
09Which languages work?
Use English for the documented examples. Validate other languages on representative inputs before relying on the results.
10How do I start?
Create an account and an API key in the console, read the quickstart, try requests in the playground, or point your AI coding agent at llms.txt.

GET STARTED

Start building.

Bring your next idea to life with Vansa. Create an account and make your first API call.

Create an account Try the playground

  1. Step 1: Create an account

    Sign up with your email, Google or GitHub.

  2. Step 2: Create an API key

    Generate a key in your console. Keep it on your server.

  3. Step 3: Add tokens

    Plans from $5/month or $1 prepaid. No output fees.

  4. Step 4: Make your first call

    Follow the quickstart to connect your application.

Read the quickstart