Skip to content
Avalanche Builders Program

Build on Avalanche with every AI model

Teams building on Avalanche can apply for a discount package on the models they use, with one API key for every text, image, video and audio model.

What you get

One key for every model your product needs

One OpenAI-compatible key

Use the tools and SDKs you already have. Switch models by changing one string.

Every model

The full catalog: 264 text, image, video, audio and embedding models, listed with their prices on the Models page.

Prepaid, no subscription

Top up by card or in USDC and pay per request at each model's list price, less your discount.

A model discount package

Approved Avalanche builders get a discount on the models they use, set per model on their account after review.

Use cases

What teams build with it

Buyback

Buy back your own token as you use AI

Every account picks a buyback token, and a share of each spend buys it back.

A share of every request you pay for (0.5% at Noob, 2% at Pro, 5% at Elite) is taken from our margin, never added to your bill, and used to buy your chosen token in public batches. $INFX is the default. Custom-token buybacks are marked coming soon in the product: until a token is tradable its share accrues in USD.

On Avalanche. Buybacks are bought on Solana, so the token has to be an eligible Solana token: a token that exists only on Avalanche cannot be picked. $INFX, which debuts on Avalanche, Base and Solana, is the default.

Every batch is listed with its transaction on the proof page.

Build

Power your AI app

One OpenAI-compatible key for every model.

Chat, image, video, audio and embedding models on one base URL, with ids like provider/model. The same key also speaks the Anthropic Messages, OpenAI Responses and Gemini API formats. Prepaid, pay per request at each model's list price.

Create

Generate images and video for content marketing

Make campaign visuals and clips through the same key.

Image and video models from several makers, billed per output unit (an image, a second of video) and priced on each model's page. Media endpoints are in preview, and generated files are deleted after 24 hours, so download what you keep.

Agents

Let your agents use paid tools

Install marketplace tools, or pay per call with x402.

Tools you install from the marketplace appear in your MCP connection, and each call is charged to your balance at its listed price. Agents that already have a wallet provider with spending policies can pay per request over x402 in USDC on Solana, for image, video and audio generation.

Rewards

Earn cash back on usage

Up to 5% of what you spend comes back to you.

Everyone pays the same list price and gets 0.5% back. Pro (1,337 $INFX on a linked Solana wallet) gets 2%, and Elite (also 1,337 $XAVA on a linked Avalanche wallet) gets 5%, paid in USDC, $INFX or another supported token, out of our margin. Holding $INFX also earns AI credits funded by LP fees, which can be small or zero.

Search and voice

Add search and speech

Embeddings and voice models on the same key.

Embedding models through POST /v1/embeddings for search and retrieval. Text-to-speech and speech-to-text models sit in the audio category, billed per character or minute of audio.

Case studies

Real use, and examples for Avalanche

Example playbooks

These are examples, not customers. Each describes how a team on Avalanche could use the product. They are not real companies, and carry no results or quotes.

Example playbook

An in-app assistant with paid tools

A wallet or consumer app on Avalanche.

A chat model answers questions inside the app through the OpenAI-compatible endpoint. The team's agents pay per request over x402 for the images and audio they generate, and call marketplace tools billed to the account balance.

How it works

From sign-up to discounts in four steps

  1. Sign up to Xava Inference with Google, email, X, a passkey or a wallet, or sign in if you already have an account.
  2. Enter your Builder Code from the Avalanche team. Each code works for the Avalanche program only.
  3. Apply in 2 minutes: who you are, the models you use and what you spend on inference today.
  4. Get a model discount package after review. Your status shows on this page.

No Builder Code yet? Ask the Avalanche ecosystem team for one. Discounts are set per model after review and may carry an end date.