---
title: Lepton AI
description: Use Lepton AI's serverless GPU inference with Nexus.
---

Lepton AI provides serverless GPU inference for open-source models. Uses an OpenAI-compatible API.

## Installation

```go
import "github.com/xraph/nexus/providers/lepton"
```

## Quick Start

```go
provider := lepton.New(os.Getenv("LEPTON_API_KEY"))

gw := nexus.New(
    nexus.WithProvider(provider),
)
```

## Options

| Option | Description |
|--------|-------------|
| `lepton.WithBaseURL(url)` | Override the API base URL (default: `https://api.lepton.ai/v1`) |

## Capabilities

| Capability | Supported |
|-----------|-----------|
| Chat | Yes |
| Streaming | Yes |
| Embeddings | No |
| Vision | No |
| Tools | No |
| Thinking | No |

## Models

| Model | Context | Max Output | Input Price | Output Price |
|-------|---------|-----------|-------------|-------------|
| `llama3.1-8b` | 131K | 4,096 | $0.07/M | $0.07/M |
| `llama3.1-70b` | 131K | 4,096 | $0.80/M | $0.80/M |