# Outrun AI

> Outrun is a prepaid AI API. One OpenAI-compatible API key reaches 86 models (100 model IDs including alternative routes), 30 uncensored, across text, image, video and speech: GPT, Claude, Gemini, Grok, DeepSeek, GLM, Kimi, MiniMax and more. Every request is billed per use from one USD balance at a published rate.

Key facts:

- API base URL: https://api.outrun.to/v1 (OpenAI Chat Completions and Responses, Anthropic Messages and Gemini native formats)
- One key and one balance for every model; no separate accounts for GPT, Claude, Gemini or others
- Prices are published per 1M tokens; the lowest text input price is $0.02. Live price list as JSON: https://api.outrun.to/pricing
- Privacy: Outrun does not store your prompts, images or outputs unless you choose to save them. API calls are not saved to chat history
- The model ID you request is the model that answers; no silent swaps to cheaper models
- Eligible new accounts get $0.10 in free API credit
- Works with Codex CLI, Claude Code, Gemini CLI, Cline, OpenCode, Grok Build CLI, Roo Code and Zo Computer
- Also available as a Telegram bot: https://t.me/outrunAPI_Bot
- Support: https://t.me/outrunsupport
- Cheapest text price: Claude Haiku 5.5 (2) (`claude-haiku-5-5-2`), $0.02 in / $0.10 out per 1M tokens

## When to use Outrun

Best-fit jobs:

- Uncensored and abliterated models on a normal OpenAI-compatible key: 30 of the 86 models, each with a published price (https://outrun.to/uncensored-ai-api)
- Prepaid with crypto or a prepaid code, no subscription: one USD balance pays for every model
- Claude Code, Codex CLI and Gemini CLI on one key and one balance: https://outrun.to/docs/setup/claude, https://outrun.to/docs/setup/codex and https://outrun.to/docs/setup/gemini
- Image and video models on the same balance as the text models

Not a fit: card billing. Outrun does not bill cards; top up with crypto or redeem a prepaid code.

## Product pages

- [Home](https://outrun.to/): overview, example request and model catalog
- [About Outrun](https://outrun.to/about): what Outrun is, how prepaid billing and crypto top-ups work, data handling and how to get support
- [Discounted AI API pricing](https://outrun.to/discounted-ai-api): Outrun per-token prices compared with official list prices for GPT, Claude, Gemini and Grok, with a cost calculator
- [Uncensored AI API](https://outrun.to/uncensored-ai-api): uncensored and abliterated LLMs plus uncensored image and video models, each with a price
- [DeepSeek API](https://outrun.to/deepseek-api): DeepSeek V4.1 Flash, V4 Pro and uncensored DeepSeek through one OpenAI-compatible key. V4.1 Flash from $0.30 in, routes from $0.20
- [GLM API](https://outrun.to/glm-api): GLM-5.3, GLM-5.3 Flash and uncensored GLM through one OpenAI-compatible key
- [GPT API](https://outrun.to/gpt-api): GPT-6.1 Sol, GPT-6 Sol, GPT-6 Astra and GPT-5.6 through one OpenAI-compatible key, with Chat Completions and Responses
- [Claude API](https://outrun.to/claude-api): Claude Sonnet, Opus and Fable through one key that works with the Anthropic Messages API and OpenAI Chat Completions. Sonnet 5.5 from $0.49 in, routes from $0.40
- [Gemini API](https://outrun.to/gemini-api): Gemini 3.8 Flash, Gemini 3.7 Flash, Gemini 4 Argon, Gemini 3.1 Pro Preview through one OpenAI-compatible key
- [Grok API](https://outrun.to/grok-api): Grok 4.7 through one OpenAI-compatible key
- [Kimi API](https://outrun.to/kimi-api): Kimi K3.1 Pro, Kimi K3 through one OpenAI-compatible key
- [MiniMax API](https://outrun.to/minimax-api): MiniMax M3.1 Flash, MiniMax M3, MiniMax M3 Uncensored through one OpenAI-compatible key
- [Qwen API](https://outrun.to/qwen-api): Qwen 3.8 Flash, Qwen 3.7 Plus, Qwen 3.7 Max, Qwen 3.8 Max through one OpenAI-compatible key

## Model pages

- [All models](https://outrun.to/model/): every model with its billed price, routes and setup values
- [GPT-6 Astra](https://outrun.to/model/gpt-6-astra): `gpt-6-astra`, $0.59 in / $2.94 out per 1M tokens
- [GPT-5.6 Sol](https://outrun.to/model/gpt-5.6-sol): `gpt-5.6-sol`, $0.30 in / $1.77 out per 1M tokens
- [GPT-5.6 Terra](https://outrun.to/model/gpt-5.6-terra): `gpt-5.6-terra`, $0.12 in / $0.71 out per 1M tokens
- [GPT-6 Sol](https://outrun.to/model/gpt-6-sol): `gpt-6-sol`, $0.12 in / $0.59 out per 1M tokens
- [GPT-6.1 Sol](https://outrun.to/model/gpt-6.1-sol): `gpt-6.1-sol`, $0.12 in / $0.59 out per 1M tokens
- [Claude Opus 5.5](https://outrun.to/model/claude-opus-5-5): `claude-opus-5-5` + 1 route IDs, $2.24 in / $11.20 out per 1M tokens; cheapest route `claude-opus-5-5-2` $0.80 in / $4.00 out
- [Claude Opus 5](https://outrun.to/model/claude-opus-5): `claude-opus-5` + 1 route IDs, $3.15 in / $15.75 out per 1M tokens; cheapest route `claude-opus-5-2` $1.00 in / $5.00 out
- [Claude Sonnet 5](https://outrun.to/model/claude-sonnet-5): `claude-sonnet-5` + 1 route IDs, $1.12 in / $5.60 out per 1M tokens; cheapest route `claude-sonnet-5-2` $0.40 in / $2.00 out
- [Claude Fable 5.1](https://outrun.to/model/claude-fable-5-1): `claude-fable-5-1` + 1 route IDs, $6.30 in / $31.50 out per 1M tokens; cheapest route `claude-fable-5-1-2` $2.00 in / $10.00 out
- [Claude Fable 5](https://outrun.to/model/claude-fable-5): `claude-fable-5` + 1 route IDs, $5.60 in / $28.00 out per 1M tokens; cheapest route `claude-fable-5-2` $2.00 in / $10.00 out
- [Claude Haiku 5.5](https://outrun.to/model/claude-haiku-5-5): `claude-haiku-5-5` + 1 route IDs, $0.09 in / $0.42 out per 1M tokens; cheapest route `claude-haiku-5-5-2` $0.02 in / $0.10 out
- [Gemini 3.8 Flash](https://outrun.to/model/gemini-3.8-flash): `gemini-3.8-flash`, $0.32 in / $1.58 out per 1M tokens
- [Grok 4.7](https://outrun.to/model/grok-4.7): `grok-4.7`, $0.09 in / $0.26 out per 1M tokens
- [Nano Banana Pro / Gemini 3 Pro Image](https://outrun.to/model/gemini-3-pro-image): `gemini-3-pro-image`, $0.04 per image
- [Nano Banana / Gemini 3.1 Flash Image](https://outrun.to/model/gemini-3.1-flash-image): `gemini-3.1-flash-image`, $0.03 per image
- [GPT Image 2](https://outrun.to/model/gpt-image-2): `gpt-image-2`, $0.06 per image
- [Grok Imagine Image](https://outrun.to/model/grok-imagine-image): `grok-imagine-image`, $0.06 per image
- [Gemini 3 STT](https://outrun.to/model/gemini-3-stt): `gemini-3-stt`, $0.32 in / $1.89 out per 1M tokens
- [Gemini 3.1 TTS](https://outrun.to/model/gemini-3.1-tts): `gemini-3.1-tts`, $0.04 per request
- [Gemini 3.8 Flash TTS](https://outrun.to/model/gemini-3.8-flash-tts): `gemini-3.8-flash-tts`, $0.32 in / $5.67 out per 1M tokens
- [Grok TTS](https://outrun.to/model/grok-tts): `grok-tts`, $9.45 per 1M characters
- [Gemini Omni Flash](https://outrun.to/model/gemini-omni-flash): `gemini-omni-flash`, $0.06 per second
- [Veo 3.1](https://outrun.to/model/veo-3.1): `veo-3.1`, $0.08 per second
- [Veo 3.1 Fast](https://outrun.to/model/veo-3.1-fast): `veo-3.1-fast`, $0.03 per second
- [Seedance 2.5 720p](https://outrun.to/model/seedance-2.5-720p): `seedance-2.5-720p`, $0.18 per second
- [Seedance 2.5 1080p](https://outrun.to/model/seedance-2.5-1080p): `seedance-2.5-1080p`, $0.47 per second
- [Seedance 2.0 720p](https://outrun.to/model/seedance-2.0-720p): `seedance-2.0-720p`, $0.13 per second
- [Seedance 2.0 1080p](https://outrun.to/model/seedance-2.0-1080p): `seedance-2.0-1080p`, $0.32 per second
- [Grok Imagine Video 1.5 1080p](https://outrun.to/model/grok-imagine-video-1.5-1080p): `grok-imagine-video-1.5-1080p`, $0.09 per second
- [Wan 3.0 1080p](https://outrun.to/model/wan3.0-video-1080p): `wan3.0-video-1080p`, $0.33 per second
- [Grok Imagine Video](https://outrun.to/model/grok-imagine-video): `grok-imagine-video`, $0.03 per second
- [DeepSeek V4 Pro](https://outrun.to/model/deepseek-v4-pro): `deepseek-v4-pro`, $1.35 in / $4.05 out per 1M tokens
- [DeepSeek V4 Flash](https://outrun.to/model/deepseek-v4-flash): `deepseek-v4-flash`, $0.45 in / $1.35 out per 1M tokens
- [DeepSeek V4.1 Flash](https://outrun.to/model/deepseek-v4.1-flash): `deepseek-v4.1-flash`, $0.30 in / $1.20 out per 1M tokens
- [GLM-5.3 Flash](https://outrun.to/model/glm-5.3-flash): `glm-5.3-flash`, $0.12 in / $0.42 out per 1M tokens
- [Claude Sonnet 5.5](https://outrun.to/model/claude-sonnet-5-5): `claude-sonnet-5-5` + 1 route IDs, $0.49 in / $2.45 out per 1M tokens; cheapest route `claude-sonnet-5-5-2` $0.40 in / $2.00 out
- [GLM-5.3 Flash Uncensored](https://outrun.to/model/glm-5.3-flash-uncensored): `glm-5.3-flash-uncensored` + 2 route IDs, $0.05 in / $0.05 out per 1M tokens
- [GLM-5.3 Uncensored](https://outrun.to/model/glm-5.3-uncensored): `glm-5.3-uncensored` + 2 route IDs, $0.10 in / $0.10 out per 1M tokens
- [DeepSeek V4 Pro Uncensored](https://outrun.to/model/deepseek-v4-pro-uncensored): `deepseek-v4-pro-uncensored`, $0.50 in / $0.50 out per 1M tokens
- [GLM-5.3 Abliterated](https://outrun.to/model/glm-5.3-abliterated): `glm-5.3-abliterated` + 1 route IDs, $1.40 in / $1.40 out per 1M tokens
- [DeepSeek V4.1 Flash Abliterated Fast](https://outrun.to/model/deepseek-v4.1-flash-abliterated-fast): `deepseek-v4.1-flash-abliterated-fast`, $1.40 in / $1.40 out per 1M tokens
- [MiMo V2.6 Flash Uncensored Thinking](https://outrun.to/model/mimo-v2.6-flash-uncensored-thinking): `mimo-v2.6-flash-uncensored-thinking`, $1.00 in / $3.00 out per 1M tokens
- [MiMo V2.6 Flash Abliterated](https://outrun.to/model/mimo-v2.6-flash-abliterated): `mimo-v2.6-flash-abliterated`, $1.00 in / $3.00 out per 1M tokens
- [MiMo V2.6 Flash Uncensored](https://outrun.to/model/mimo-v2.6-flash-uncensored): `mimo-v2.6-flash-uncensored`, $1.00 in / $3.00 out per 1M tokens
- [Kimi K3](https://outrun.to/model/kimi-k3): `kimi-k3`, $1.80 in / $9.00 out per 1M tokens
- [GLM-5.3](https://outrun.to/model/glm-5.3): `glm-5.3`, $0.66 in / $2.06 out per 1M tokens
- [MiniMax M3](https://outrun.to/model/minimax-m3): `MiniMax-M3`, $0.18 in / $0.72 out per 1M tokens
- [GLM-5.3 Flash Abliterated](https://outrun.to/model/glm-5.3-flash-abliterated): `glm-5.3-flash-abliterated`, $1.32 in / $1.32 out per 1M tokens
- [DeepSeek V4.1 Flash 1017](https://outrun.to/model/deepseek-v4.1-flash-1017): `deepseek-v4.1-flash-1017-2` + 1 route IDs, $0.39 in / $2.33 out per 1M tokens; cheapest route `deepseek-v4.1-flash-1017-3` $0.20 in / $1.20 out
- [DeepSeek V4.1 Flash Abliterated](https://outrun.to/model/deepseek-v4.1-flash-abliterated): `deepseek-v4.1-flash-abliterated` + 1 route IDs, $2.10 in / $2.10 out per 1M tokens; cheapest route `deepseek-v4.1-flash-abliterated-2` $0.30 in / $0.30 out
- [GLM-5.3 Flash Code Uncensored](https://outrun.to/model/glm-5.3-flash-code-uncensored): `glm-5.3-flash-code-uncensored`, $12.00 in / $12.00 out per 1M tokens
- [GPT OSS 120B Uncensored](https://outrun.to/model/gpt-oss-120b-uncensored): `gpt-oss-120b-uncensored`, $0.22 in / $1.46 out per 1M tokens
- [MiniMax M3 Uncensored](https://outrun.to/model/minimax-m3-uncensored): `minimax-m3-uncensored`, $1.50 in / $8.00 out per 1M tokens
- [Outrun Red Team Flash](https://outrun.to/model/outrun-redteam-flash): `outrun-redteam-flash`, $1.35 in / $4.50 out per 1M tokens
- [Outrun Red Team Pro](https://outrun.to/model/outrun-redteam-pro): `outrun-redteam-pro`, $2.25 in / $9.00 out per 1M tokens
- [Hermes 4 405B](https://outrun.to/model/hermes-4-405b-uncensored): `hermes-4-405b-uncensored`, $1.80 in / $7.20 out per 1M tokens
- [Nemotron 3 Ultra 550B Uncensored](https://outrun.to/model/nemotron-3-ultra-550b-a55b-uncensored): `nemotron-3-ultra-550b-a55b-uncensored`, $1.69 in / $9.90 out per 1M tokens
- [GLM-5.3 Flash Cybersecurity](https://outrun.to/model/glm-5.3-flash-cybersecurity): `glm-5.3-flash-cybersecurity`, $0.30 in / $1.00 out per 1M tokens
- [DeepSeek V4.1 Flash 1017 Thinking](https://outrun.to/model/deepseek-v4.1-flash-1017-thinking): `deepseek-v4.1-flash-1017-thinking-3`, $0.20 in / $1.20 out per 1M tokens
- [DeepSeek V4.1 Flash Uncensored 1017](https://outrun.to/model/deepseek-v4.1-flash-1017-uncensored): `deepseek-v4.1-flash-1017-uncensored-3`, $0.66 in / $2.64 out per 1M tokens
- [DeepSeek V4.1 Flash Uncensored Thinking 1017](https://outrun.to/model/deepseek-v4.1-flash-1017-uncensored-thinking): `deepseek-v4.1-flash-1017-uncensored-thinking-3`, $0.66 in / $2.64 out per 1M tokens
- [Apex Flash 1 Abliterated](https://outrun.to/model/glm-5.3-flash-apex-abliterated): `glm-5.3-flash-apex-abliterated`, $0.30 in / $1.00 out per 1M tokens
- [Anime Image Uncensored](https://outrun.to/model/anime-image-uncensored): `anime-image-uncensored`, $0.06 per image
- [Chroma Uncensored](https://outrun.to/model/chroma-uncensored): `chroma-uncensored`, $0.16 per image
- [Lustify SDXL Uncensored](https://outrun.to/model/lustify-sdxl-uncensored): `lustify-sdxl-uncensored`, $0.07 per image
- [Seedream 5.0 Lite Edit Uncensored](https://outrun.to/model/seedream-5.0-lite-edit-uncensored): `seedream-5.0-lite-edit-uncensored`, $0.07 per image
- [Vision Image Uncensored](https://outrun.to/model/vision-image-uncensored): `vision-image-uncensored`, $0.08 per image
- [Vision Pro Image Uncensored](https://outrun.to/model/vision-pro-image-uncensored): `vision-pro-image-uncensored`, $0.19 per image
- [CyberRealistic XL Uncensored](https://outrun.to/model/cyberrealistic-xl-uncensored): `cyberrealistic-xl-uncensored`, $0.06 per image
- [WAI Illustrious SDXL Uncensored](https://outrun.to/model/illustrious-xl-uncensored): `illustrious-xl-uncensored`, $0.06 per image
- [Qwen Image 2.1 Uncensored](https://outrun.to/model/qwen-image-2.1-uncensored): `qwen-image-2.1-uncensored`, $0.08 per image
- [MiniMax H3](https://outrun.to/model/minimax-h3): `MiniMax-H3`, $0.06 per second
- [Wan 3.0 Video Uncensored](https://outrun.to/model/wan-3.0-video-uncensored): `wan-3.0-video-uncensored`, $0.40 per second
- [Mercury 3](https://outrun.to/model/mercury-3): `mercury-3`, $0.20 in / $0.70 out per 1M tokens
- [Gemini 4 Argon](https://outrun.to/model/gemini-4-argon): `gemini-4-argon`, $1.20 in / $6.00 out per 1M tokens
- [MiniMax M3.1 Flash](https://outrun.to/model/minimax-m3.1-flash): `minimax-m3.1-flash`, $0.09 in / $0.36 out per 1M tokens
- [Gemini 3.7 Flash](https://outrun.to/model/gemini-3.7-flash): `gemini-3.7-flash`, $0.95 in / $4.73 out per 1M tokens
- [Gemini 3.1 Pro Preview](https://outrun.to/model/gemini-3.1-pro-preview): `gemini-3.1-pro-preview`, $1.26 in / $7.56 out per 1M tokens
- [Kimi K3.1 Pro](https://outrun.to/model/kimi-k3.1-pro): `kimi-k3.1-pro`, $0.20 in / $0.20 out per 1M tokens
- [Qwen 3.7 Max](https://outrun.to/model/qwen3.7-max): `qwen3.7-max`, $1.44 in / $4.32 out per 1M tokens
- [Qwen 3.8 Flash](https://outrun.to/model/qwen3.8-flash): `qwen3.8-flash`, $0.10 in / $0.33 out per 1M tokens
- [Qwen 3.8 Max](https://outrun.to/model/qwen3.8-max): `qwen3.8-max`, $1.44 in / $4.32 out per 1M tokens
- [Qwen 3.7 Plus](https://outrun.to/model/qwen3.7-plus): `qwen3.7-plus`, $0.29 in / $1.14 out per 1M tokens
- [GPT Image 2.5 Flare](https://outrun.to/model/gpt-image-2.5-flare): `gpt-image-2.5-flare`, $0.02 per image
- [GPT Image 2.5 Sunburst](https://outrun.to/model/gpt-image-2.5-sunburst): `gpt-image-2.5-sunburst`, $0.02 per image
- [Seedream 5 Pro](https://outrun.to/model/doubao-seedream-5-0-pro): `doubao-seedream-5-0-pro`, $0.05 per image

## Docs

- [Full text: this file, the docs and the price list](https://outrun.to/llms-full.txt)
- [OpenAPI 3.1 description](https://outrun.to/openapi.json)
- Markdown copies: add .md to a page URL, for example https://outrun.to/claude-api.md, or send Accept: text/markdown
- [Documentation home](https://outrun.to/docs/)
- [API reference](https://outrun.to/docs/api.md): endpoints, authentication and request formats
- [Models](https://outrun.to/docs/models): current model catalog
- [Billing](https://outrun.to/docs/billing): prepaid balance and per-use charges
- [Playground](https://outrun.to/docs/playground)
- [Troubleshooting](https://outrun.to/docs/troubleshooting)
- [Support](https://outrun.to/docs/support)

## Setup guides

- [SillyTavern](https://outrun.to/docs/setup/sillytavern)
- [LiteLLM](https://outrun.to/docs/setup/litellm)

- [Claude Code](https://outrun.to/docs/setup/claude.md)
- [Codex CLI](https://outrun.to/docs/setup/codex.md)
- [Codex desktop app](https://outrun.to/docs/setup/codex_app)
- [Gemini CLI](https://outrun.to/docs/setup/gemini)
- [OpenCode CLI](https://outrun.to/docs/setup/opencode)
- [Cline (VS Code)](https://outrun.to/docs/setup/cline)
- [Grok Build CLI](https://outrun.to/docs/setup/grokbuild)
- [Zo Computer](https://outrun.to/docs/setup/zo)
- [Uncensored models in Roo Code](https://outrun.to/docs/setup/uncensored)
- [Uncensored on iPhone and Android](https://outrun.to/docs/setup/uncensored_mobile)
- [Image generation](https://outrun.to/docs/setup/images)

## Optional

- [Privacy policy](https://outrun.to/privacy)
- [Terms of service](https://outrun.to/terms)

---

# Outrun documentation, full text

Copied from the published Markdown versions of the docs pages. Other setup guides are linked in the lists above.

============================================================
Source: https://outrun.to/docs/api (API reference)
============================================================

# Outrun API Reference — Endpoints & Authentication

> Explore Outrun API endpoints, authentication, supported protocols and request examples for chat, images, video and account balance.

Canonical: https://outrun.to/docs/api

OpenAI, Anthropic and Gemini compatible endpoints. One Outrun key (sk-rs-...) works for every endpoint and every model.

## Base URLs

Enter the base URL in your app exactly as shown. Apps add the request path (/chat/completions, /messages, ...) themselves.

| Protocol | Base URL | Used by |
| --- | --- | --- |
| OpenAI-compatible | https://api.outrun.to/v1 | OpenAI SDKs, Codex, OpenCode, Cline, Zo Computer, Grok Build, most apps |
| Anthropic Messages | https://api.outrun.to | Claude Code, Anthropic SDKs (bare host, no /v1) |
| Gemini native | https://api.outrun.to/v1/gemini-cli | Gemini CLI, Google Gen AI SDKs (the client adds /v1beta/...) |

If an OpenAI-compatible app adds /v1 by itself, use the bare host https://api.outrun.to instead. Gemini native requests must go through https://api.outrun.to/v1/gemini-cli; plain https://api.outrun.to/v1beta/... is blocked by the firewall (403).

## Authentication

Get your key in @outrunAPI_Bot -> My API key. Send it in one of these headers; never share it or commit it to a repository.

| Header | Example | Typical use |
| --- | --- | --- |
| Authorization | Authorization: Bearer $OUTRUN_API_KEY | OpenAI-style requests (all /v1 endpoints) |
| x-api-key | x-api-key: $OUTRUN_API_KEY (plus anthropic-version: 2023-06-01) | Anthropic-style requests (/v1/messages) |
| x-goog-api-key or ?key= | x-goog-api-key: $OUTRUN_API_KEY | Gemini native requests (/v1/gemini-cli/v1beta/...) |

## Endpoints

| Method | Path | What it does | Models |
| --- | --- | --- | --- |
| POST | / v1/ chat/ completions | OpenAI Chat Completions (add "stream": true to stream) | All text models; image models for chat-to-image |
| POST | / v1/ responses | OpenAI Responses API | GPT models (used by Codex) |
| POST | / v1/ messages | Anthropic Messages API | Claude models (used by Claude Code) |
| POST | / v1/ messages/ count_tokens | Anthropic token counting | Claude models |
| GET | / v1/ models | List the model IDs available to your key | - |
| POST | / v1/ images/ generations | Generate an image | gemini-3-pro-image , gemini-3.1-flash-image , gpt-image-2 , grok-imagine-image |
| POST | / v1/ audio/ speech | Text to speech | gemini-3.1-tts, gemini-3.8-flash-tts, grok-tts |
| POST | / v1/ chat/ completions (input_audio) | Speech to text: send the audio inside a chat message | gemini-3-stt |
| POST | / v1/ videos | Start a video generation task | Video models |
| GET | / v1/ videos/ {id} | Check a video task | Video models |
| GET | / v1/ videos/ {id}/ content | Download the finished video (mp4) | Video models |
| POST | / v1/ gemini-cli/ v1beta/ models/ {model}: generateContent | Gemini native generation | Gemini models |
| POST | / v1/ gemini-cli/ v1beta/ models/ {model}: streamGenerateContent | Gemini native streaming (add ?alt=sse) | Gemini models |
| GET | / v1/ account | Your Outrun Credit balance in USD | - |
| GET | / pricing | Public price list as JSON (no key needed) | - |

Model IDs are case-sensitive and must match Models & Pricing exactly. /v1/audio/transcriptions is not available; use the input_audio example below for speech to text.

## Chat Completions: request and response

cURL Request

Request · POST /v1/chat/completions

```
curl https://api.outrun.to/v1/chat/completions \
  -H "Authorization: Bearer $OUTRUN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.6-sol",
    "messages": [
      {"role": "system", "content": "You are a helpful assistant."},
      {"role": "user", "content": "Say hello in five words."}
    ],
    "max_tokens": 64
  }'
```

JSON Response (200 OK)

Response · application/json

```
{
  "id": "chatcmpl-...",
  "object": "chat.completion",
  "model": "gpt-5.6-sol",
  "choices": [
    {
      "index": 0,
      "message": { "role": "assistant", "content": "Hello there, nice to meet!" },
      "finish_reason": "stop"
    }
  ],
  "usage": { "prompt_tokens": 24, "completion_tokens": 8, "total_tokens": 32 }
}
```

## Responses API (GPT, Codex)

POST /v1/responses

```
curl https://api.outrun.to/v1/responses \
  -H "Authorization: Bearer $OUTRUN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "gpt-5.6-sol", "input": "Write a haiku about rivers.", "reasoning": {"effort": "high"}}'
```

## Anthropic Messages (Claude)

POST /v1/messages

```
curl https://api.outrun.to/v1/messages \
  -H "x-api-key: $OUTRUN_API_KEY" \
  -H "anthropic-version: 2023-06-01" \
  -H "Content-Type: application/json" \
  -d '{"model": "claude-sonnet-5", "max_tokens": 256, "messages": [{"role": "user", "content": "Hello, Claude"}]}'

# Count tokens before sending:
curl https://api.outrun.to/v1/messages/count_tokens \
  -H "x-api-key: $OUTRUN_API_KEY" \
  -H "anthropic-version: 2023-06-01" \
  -H "Content-Type: application/json" \
  -d '{"model": "claude-sonnet-5", "messages": [{"role": "user", "content": "Hello, Claude"}]}'
```

## Gemini native

POST /v1/gemini-cli/v1beta/models/{model}:generateContent

```
curl "https://api.outrun.to/v1/gemini-cli/v1beta/models/gemini-3.8-flash:generateContent" \
  -H "x-goog-api-key: $OUTRUN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"contents": [{"role": "user", "parts": [{"text": "Explain DNS in one sentence."}]}]}'

# Streaming (server-sent events); the key can also go in ?key=
curl "https://api.outrun.to/v1/gemini-cli/v1beta/models/gemini-3.8-flash:streamGenerateContent?alt=sse&key=$OUTRUN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"contents": [{"role": "user", "parts": [{"text": "Count to five."}]}]}'
```

## List models

GET /v1/models

```
curl https://api.outrun.to/v1/models \
  -H "Authorization: Bearer $OUTRUN_API_KEY"
```

## Image generation

POST /v1/images/generations

```
curl https://api.outrun.to/v1/images/generations \
  -H "Authorization: Bearer $OUTRUN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "gemini-3-pro-image", "prompt": "A lighthouse at dawn, oil painting", "size": "16:9"}'

# Response: {"created": 1790000000, "data": [{"b64_json": "<base64 PNG>"}], "usage": {...}}
# Gemini image models return one image per request (n = 1).
# gpt-image-2 and grok-imagine-image use the same request shape.
```

## Text to speech

POST /v1/audio/speech

```
curl https://api.outrun.to/v1/audio/speech \
  -H "Authorization: Bearer $OUTRUN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "gemini-3.1-tts", "input": "Welcome to Outrun.", "voice": "Kore", "response_format": "wav"}' \
  --output speech.wav

# Use a Gemini voice name (for example Kore) and response_format wav or pcm.
# gemini-3.8-flash-tts: the same Gemini voices; wav, mp3, flac or pcm; inline tags such as <laugh>.
# grok-tts: a Grok preset voice (default eve); mp3, wav, pcm, mulaw or alaw; speed 0.7 to 1.5.
# Input is limited to 8,000 bytes per request. Custom or cloned voices are not available.
```

## Speech to text

POST /v1/chat/completions with input_audio

```
# macOS / Linux: put the audio file into a JSON body as base64
B64=$(base64 < clip.mp3 | tr -d '\n')
cat > stt.json <<EOF
{"model": "gemini-3-stt",
 "messages": [{"role": "user", "content": [
   {"type": "text", "text": "Transcribe this audio."},
   {"type": "input_audio", "input_audio": {"data": "$B64", "format": "mp3"}}
 ]}]}
EOF
curl https://api.outrun.to/v1/chat/completions \
  -H "Authorization: Bearer $OUTRUN_API_KEY" \
  -H "Content-Type: application/json" \
  -d @stt.json
```

## Video generation

POST /v1/videos, then GET /v1/videos/{id}

```
# 1) Start a task
curl https://api.outrun.to/v1/videos \
  -H "Authorization: Bearer $OUTRUN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "veo-3.1-fast", "prompt": "A paper boat drifting down a rainy street", "seconds": 4, "size": "16:9"}'
# -> returns a task id and "status": "queued"

# 2) Check it every few seconds until "status" is "completed" (or "failed")
curl https://api.outrun.to/v1/videos/VIDEO_ID \
  -H "Authorization: Bearer $OUTRUN_API_KEY"

# 3) Download the mp4
curl -L https://api.outrun.to/v1/videos/VIDEO_ID/content \
  -H "Authorization: Bearer $OUTRUN_API_KEY" \
  -o video.mp4
```

[Try video and image models in the Playground](https://outrun.to/docs/playground)

Video is billed per second of output. Allowed lengths depend on the model; veo-3.1, veo-3.1-fast and gemini-omni-flash accept size 16:9 or 9:16; add "resolution": "4k" on veo-3.1 or veo-3.1-fast for the 4K tier. Links to finished videos point to api.outrun.to.

| Model | Allowed seconds |
| --- | --- |
| veo-3.1 , veo-3.1-fast | 4, 6 or 8 |
| gemini-omni-flash | 3 to 10 |
| seedance-2.5-720p , seedance-2.5-1080p , wan3.0-video-1080p | 4 to 30 |
| seedance-2.0-720p , seedance-2.0-1080p | 4 to 15 |
| grok-imagine-video , grok-imagine-video-1.5-1080p | 1 to 15 |

## Check your credit

GET /v1/account

```
curl https://api.outrun.to/v1/account \
  -H "Authorization: Bearer $OUTRUN_API_KEY"

# Response: {"currency": "USD", "credit": 12.34}
```

## Errors

Errors use the usual OpenAI / Anthropic JSON error shape. See Troubleshooting for fixes.

| Status | Meaning |
| --- | --- |
| 401 | "Invalid token" or "Invalid API key": the key is missing, mistyped or was replaced |
| Insufficient balance | Not enough Outrun Credit for the request; redeem a new code in the bot |
| 429 | Rate limit; wait a moment and retry |
| 503 | "<model> is temporarily paused"; try another model or check /status in the bot |
| 5xx | Temporary server error; retry after a short wait |

============================================================
Source: https://outrun.to/docs/setup/claude (Claude Code setup)
============================================================

# Outrun API Setup for Claude Code

> Connect Claude Code to Outrun using the documented endpoint and authentication settings. Follow setup steps and model configuration examples.

Canonical: https://outrun.to/docs/setup/claude



## Windows

Follow these steps in a terminal window. Choose your computer below.

### Before you start

Have your API key and some Outrun credit ready. Keep your key private.

[Get your API key](https://outrun.to/api-keys)

[Check balance](https://outrun.to/balance)

[Choose a model](https://outrun.to/docs/models)

Open PowerShell: open Start, type PowerShell, then open Windows PowerShell. Commands below are for this window.

[GitHub Repository](https://github.com/anthropics/claude-code)

**Optional setup helper Use an AI assistant to walk through these steps**

Copy the setup instructions into an AI assistant you already use. Review any file changes it suggests before allowing them.

Enter your API key directly in the app or local settings. Keep it out of the assistant’s chat.

**Connection settings Show ⌄**

Connection settings

```
ANTHROPIC_BASE_URL: https://api.outrun.to
Key variable: ANTHROPIC_AUTH_TOKEN
First model: claude-opus-5-5
```

Get your key from Outrun → API keys and enter it privately in the app or local settings. Test with “Reply with OK only.” Tests use credit; check Activity before retrying.

### Setup Steps

1. 1

  **Install Node.js first**

  Use the Download Node.js link below and choose the LTS version for your computer. Follow its installation instructions, then close and reopen your terminal. Node.js includes npm, the program used by the Install Command below.

  [Download Node.js (LTS)](https://nodejs.org/en/download)

2. 2

  **Install the tool**

  Open Commands and settings below. Copy only the Install Command for your selected computer into your terminal and press Enter. Wait for it to finish.

3. 3

  **Open the settings file**

  Find %USERPROFILE%\.claude\settings.json. Press Windows+R, paste the folder part of this path, and press Enter. %USERPROFILE% means your Windows user folder. Open the file in a text editor such as Notepad. If the folder or file is missing, create it.

4. 4

  **Add the Outrun settings**

  Open Commands and settings below and copy the Configuration Template into this file. If the file already has settings, keep them and add only the Outrun provider/settings block. Save the file. Do not paste JSON or TOML settings into the terminal.

5. 5

  **Enter your key privately**

  Use the Environment Setup commands for your computer below. Replace KEY with your own Outrun key in your local terminal or settings file. Never put the key in an AI chat or screenshot.

6. 6

  **Reopen and test**

  Open a new terminal, type claude, and press Enter. If you are signed in with an Anthropic account, use /logout first and accept the custom API key if asked. Start a new chat and send: Reply with OK only. A short reply with no connection error means the first test worked. This request uses Outrun credit; check Activity if it is interrupted before trying again.

**Commands and settings for Windows Show ⌄**

Install Command

```
npm install -g @anthropic-ai/claude-code
```

Windows PowerShell · user environment

```
# 1) Write the settings.json shown in the configuration template below.
# 2) Also set user environment variables (replace KEY with your own key):
[Environment]::SetEnvironmentVariable("ANTHROPIC_BASE_URL", "https://api.outrun.to", "User")
[Environment]::SetEnvironmentVariable("ANTHROPIC_AUTH_TOKEN", "KEY", "User")
[Environment]::SetEnvironmentVariable("ANTHROPIC_MODEL", "claude-opus-5-5", "User")
[Environment]::SetEnvironmentVariable("ANTHROPIC_DEFAULT_OPUS_MODEL", "claude-opus-5-5", "User")
[Environment]::SetEnvironmentVariable("ANTHROPIC_DEFAULT_SONNET_MODEL", "claude-sonnet-5-5", "User")
[Environment]::SetEnvironmentVariable("ANTHROPIC_DEFAULT_HAIKU_MODEL", "claude-sonnet-5", "User")
# Remove an older key variable (both set at once causes an auth conflict):
[Environment]::SetEnvironmentVariable("ANTHROPIC_API_KEY", $null, "User")
# Close every terminal and VS Code window, then open a new one.
```

Replace KEY with your own key only in your local app, editor or terminal. Never send that key to an AI chat. Copy one command block at a time.

### Configuration Template

Config JSON

```
// Windows: %USERPROFILE%\.claude\settings.json  (merge, keep your other settings)
{
  "model": "claude-opus-5-5",
  "env": {
    "ANTHROPIC_BASE_URL": "https://api.outrun.to",
    "ANTHROPIC_AUTH_TOKEN": "KEY",
    "ANTHROPIC_MODEL": "claude-opus-5-5",
    "ANTHROPIC_DEFAULT_OPUS_MODEL": "claude-opus-5-5",
    "ANTHROPIC_DEFAULT_SONNET_MODEL": "claude-sonnet-5-5",
    "ANTHROPIC_DEFAULT_HAIKU_MODEL": "claude-haiku-5-5"
  }
}
// Remove any "ANTHROPIC_API_KEY" entry from this file.
// All Claude models: claude-opus-5-5 (default), claude-opus-5, claude-sonnet-5-5,
// claude-sonnet-5, claude-fable-5-1, claude-fable-5. Switch with /model <model>.
```

**Optional: test with a script (uses credit) Show**

Test cURL

```
curl https://api.outrun.to/v1/messages \
  -H "x-api-key: $OUTRUN_API_KEY" \
  -H "anthropic-version: 2023-06-01" \
  -H "Content-Type: application/json" \
  -d '{"model":"claude-opus-5-5","max_tokens":16,"messages":[{"role":"user","content":"Reply with OK only."}]}' 
```

### If something goes wrong

**Key missing or rejected (401)**

Copy your current key from API keys. Paste it only into the app’s key field or your local key variable, then save and reopen the app.

**Not enough available credit (402)**

Check Balance and the key’s spending cap. Another request may have temporarily reserved credit; read the message before adding funds.

**Wrong URL, model or limit (400/404)**

Copy the exact values above. Do not add an operation such as /chat/completions to a Base URL. For GLM, context is 1000000 and maximum output is 65536.

**Busy or rate limited (429)**

Wait for the active request to finish or the rate limit to reset. Avoid repeatedly submitting the same request.

**Interrupted request or uncertain charge**

Check Activity first. Keep the request ID for support; never include your key in a message or screenshot.

**Model choices and service notes**

Claude Code uses the Anthropic Messages API on the bare host https://api.outrun.to (no /v1). All Claude models are set up: claude-opus-5-5 (default), claude-opus-5, claude-sonnet-5-5, claude-sonnet-5, claude-fable-5-1 and claude-fable-5. Each also has a lower-cost route 2: add -2 to the ID (for example claude-opus-5-5-2). If a request returns 402, check your account balance and key spending cap. For 429, read the error: wait for the active request to finish or the rate limit to reset. If a paid request is interrupted or its outcome is uncertain, check Activity before trying again; do not automatically repeat it. Choose an available model from Models & Pricing; paused models cannot be used.

[Contact @outrunsupport](https://t.me/outrunsupport)

## macOS

Follow these steps in a terminal window. Choose your computer below.

### Before you start

Have your API key and some Outrun credit ready. Keep your key private.

[Get your API key](https://outrun.to/api-keys)

[Check balance](https://outrun.to/balance)

[Choose a model](https://outrun.to/docs/models)

Open Terminal: press Command+Space, type Terminal, then press Enter. Commands below are for this window.

[**GitHub Repository](https://github.com/anthropics/claude-code)

**Optional setup helper Use an AI assistant to walk through these steps**

Copy the setup instructions into an AI assistant you already use. Review any file changes it suggests before allowing them.

Enter your API key directly in the app or local settings. Keep it out of the assistant’s chat.

**Connection settings Show ⌄**

Connection settings

```
ANTHROPIC_BASE_URL: https://api.outrun.to
Key variable: ANTHROPIC_AUTH_TOKEN
First model: claude-opus-5-5
```

Get your key from Outrun → API keys and enter it privately in the app or local settings. Test with “Reply with OK only.” Tests use credit; check Activity before retrying.

### Setup Steps

1. 1

  **Install Node.js first**

  Use the Download Node.js link below and choose the LTS version for your computer. Follow its installation instructions, then close and reopen your terminal. Node.js includes npm, the program used by the Install Command below.

  [**Download Node.js (LTS)](https://nodejs.org/en/download)

2. 2

  **Install the tool**

  Open Commands and settings below. Copy only the Install Command for your selected computer into your terminal and press Enter. Wait for it to finish.

3. 3

  **Set the connection address and key**

  In Commands and settings, use the environment setup for your shell. Replace KEY with your own key locally. ANTHROPIC_BASE_URL is https://api.outrun.to, with no /v1. Use ANTHROPIC_AUTH_TOKEN, not both token and ANTHROPIC_API_KEY.

4. 4

  **Save your shell settings**

  The commands update ~/.zshrc and ~/.bashrc. These files load your connection settings when a terminal opens. Keep unrelated lines, save, then open a new terminal.

5. 5

  **Reopen and test**

  Open a new terminal, type claude, and press Enter. If you are signed in with an Anthropic account, use /logout first and accept the custom API key if asked. Start a new chat and send: Reply with OK only. A short reply with no connection error means the first test worked. This request uses Outrun credit; check Activity if it is interrupted before trying again.

**Commands and settings for macOS Show ⌄**

Install Command

```
npm install -g @anthropic-ai/claude-code
```

macOS / Linux · ~/.zshrc and ~/.bashrc

```
# Add these lines to BOTH ~/.zshrc and ~/.bashrc (replace KEY with your own key).
# The base URL is the bare host: no /v1 at the end.
export ANTHROPIC_BASE_URL="https://api.outrun.to"
export ANTHROPIC_AUTH_TOKEN="KEY"
export ANTHROPIC_MODEL="claude-opus-5-5"
export ANTHROPIC_DEFAULT_OPUS_MODEL="claude-opus-5-5"
export ANTHROPIC_DEFAULT_SONNET_MODEL="claude-sonnet-5-5"
export ANTHROPIC_DEFAULT_HAIKU_MODEL="claude-haiku-5-5"

# Remove any older ANTHROPIC_API_KEY or ANTHROPIC_BASE_URL lines for another provider,
# then reload:
source ~/.zshrc 2>/dev/null || source ~/.bashrc
```

Replace KEY with your own key only in your local app, editor or terminal. Never send that key to an AI chat. Copy one command block at a time.

### Configuration Template

Config JSON

```
// Windows: %USERPROFILE%\.claude\settings.json  (merge, keep your other settings)
{
  "model": "claude-opus-5-5",
  "env": {
    "ANTHROPIC_BASE_URL": "https://api.outrun.to",
    "ANTHROPIC_AUTH_TOKEN": "KEY",
    "ANTHROPIC_MODEL": "claude-opus-5-5",
    "ANTHROPIC_DEFAULT_OPUS_MODEL": "claude-opus-5-5",
    "ANTHROPIC_DEFAULT_SONNET_MODEL": "claude-sonnet-5-5",
    "ANTHROPIC_DEFAULT_HAIKU_MODEL": "claude-haiku-5-5"
  }
}
// Remove any "ANTHROPIC_API_KEY" entry from this file.
// All Claude models: claude-opus-5-5 (default), claude-opus-5, claude-sonnet-5-5,
// claude-sonnet-5, claude-fable-5-1, claude-fable-5. Switch with /model <model>.
```

**Optional: test with a script (uses credit) Show**

Test cURL

```
curl https://api.outrun.to/v1/messages \
  -H "x-api-key: $OUTRUN_API_KEY" \
  -H "anthropic-version: 2023-06-01" \
  -H "Content-Type: application/json" \
  -d '{"model":"claude-opus-5-5","max_tokens":16,"messages":[{"role":"user","content":"Reply with OK only."}]}' 
```

### If something goes wrong

**Key missing or rejected (401)**

Copy your current key from API keys. Paste it only into the app’s key field or your local key variable, then save and reopen the app.

**Not enough available credit (402)**

Check Balance and the key’s spending cap. Another request may have temporarily reserved credit; read the message before adding funds.

**Wrong URL, model or limit (400/404)**

Copy the exact values above. Do not add an operation such as /chat/completions to a Base URL. For GLM, context is 1000000 and maximum output is 65536.

**Busy or rate limited (429)**

Wait for the active request to finish or the rate limit to reset. Avoid repeatedly submitting the same request.

**Interrupted request or uncertain charge**

Check Activity first. Keep the request ID for support; never include your key in a message or screenshot.

**Model choices and service notes**

Claude Code uses the Anthropic Messages API on the bare host https://api.outrun.to (no /v1). All Claude models are set up: claude-opus-5-5 (default), claude-opus-5, claude-sonnet-5-5, claude-sonnet-5, claude-fable-5-1 and claude-fable-5. Each also has a lower-cost route 2: add -2 to the ID (for example claude-opus-5-5-2). If a request returns 402, check your account balance and key spending cap. For 429, read the error: wait for the active request to finish or the rate limit to reset. If a paid request is interrupted or its outcome is uncertain, check Activity before trying again; do not automatically repeat it. Choose an available model from Models & Pricing; paused models cannot be used.

[Contact @outrunsupport](https://t.me/outrunsupport)

## Linux

Follow these steps in a terminal window. Choose your computer below.

### Before you start

Have your API key and some Outrun credit ready. Keep your key private.

[Get your API key](https://outrun.to/api-keys)

[Check balance](https://outrun.to/balance)

[Choose a model](https://outrun.to/docs/models)

Open your computer’s Terminal app. Commands below are for this window.

[GitHub Repository](https://github.com/anthropics/claude-code)

**Optional setup helper Use an AI assistant to walk through these steps**

Copy the setup instructions into an AI assistant you already use. Review any file changes it suggests before allowing them.

Enter your API key directly in the app or local settings. Keep it out of the assistant’s chat.

**Connection settings Show ⌄**

Connection settings

```
ANTHROPIC_BASE_URL: https://api.outrun.to
Key variable: ANTHROPIC_AUTH_TOKEN
First model: claude-opus-5-5
```

Get your key from Outrun → API keys and enter it privately in the app or local settings. Test with “Reply with OK only.” Tests use credit; check Activity before retrying.

### Setup Steps

1. 1

  **Install Node.js first**

  Use the Download Node.js link below and choose the LTS version for your computer. Follow its installation instructions, then close and reopen your terminal. Node.js includes npm, the program used by the Install Command below.

  [Download Node.js (LTS)](https://nodejs.org/en/download)

2. 2

  **Install the tool**

  Open Commands and settings below. Copy only the Install Command for your selected computer into your terminal and press Enter. Wait for it to finish.

3. 3

  **Set the connection address and key**

  In Commands and settings, use the environment setup for your shell. Replace KEY with your own key locally. ANTHROPIC_BASE_URL is https://api.outrun.to, with no /v1. Use ANTHROPIC_AUTH_TOKEN, not both token and ANTHROPIC_API_KEY.

4. 4

  **Save your shell settings**

  The commands update ~/.bashrc and ~/.zshrc. These files load your connection settings when a terminal opens. Keep unrelated lines, save, then open a new terminal.

5. 5

  **Reopen and test**

  Open a new terminal, type claude, and press Enter. If you are signed in with an Anthropic account, use /logout first and accept the custom API key if asked. Start a new chat and send: Reply with OK only. A short reply with no connection error means the first test worked. This request uses Outrun credit; check Activity if it is interrupted before trying again.

**Commands and settings for Linux Show ⌄**

Install Command

```
npm install -g @anthropic-ai/claude-code
```

macOS / Linux · ~/.zshrc and ~/.bashrc

```
# Add these lines to BOTH ~/.zshrc and ~/.bashrc (replace KEY with your own key).
# The base URL is the bare host: no /v1 at the end.
export ANTHROPIC_BASE_URL="https://api.outrun.to"
export ANTHROPIC_AUTH_TOKEN="KEY"
export ANTHROPIC_MODEL="claude-opus-5-5"
export ANTHROPIC_DEFAULT_OPUS_MODEL="claude-opus-5-5"
export ANTHROPIC_DEFAULT_SONNET_MODEL="claude-sonnet-5-5"
export ANTHROPIC_DEFAULT_HAIKU_MODEL="claude-haiku-5-5"

# Remove any older ANTHROPIC_API_KEY or ANTHROPIC_BASE_URL lines for another provider,
# then reload:
source ~/.zshrc 2>/dev/null || source ~/.bashrc
```

Replace KEY with your own key only in your local app, editor or terminal. Never send that key to an AI chat. Copy one command block at a time.

### Configuration Template

Config JSON

```
// Windows: %USERPROFILE%\.claude\settings.json  (merge, keep your other settings)
{
  "model": "claude-opus-5-5",
  "env": {
    "ANTHROPIC_BASE_URL": "https://api.outrun.to",
    "ANTHROPIC_AUTH_TOKEN": "KEY",
    "ANTHROPIC_MODEL": "claude-opus-5-5",
    "ANTHROPIC_DEFAULT_OPUS_MODEL": "claude-opus-5-5",
    "ANTHROPIC_DEFAULT_SONNET_MODEL": "claude-sonnet-5-5",
    "ANTHROPIC_DEFAULT_HAIKU_MODEL": "claude-haiku-5-5"
  }
}
// Remove any "ANTHROPIC_API_KEY" entry from this file.
// All Claude models: claude-opus-5-5 (default), claude-opus-5, claude-sonnet-5-5,
// claude-sonnet-5, claude-fable-5-1, claude-fable-5. Switch with /model <model>.
```

**Optional: test with a script (uses credit) Show**

Test cURL

```
curl https://api.outrun.to/v1/messages \
  -H "x-api-key: $OUTRUN_API_KEY" \
  -H "anthropic-version: 2023-06-01" \
  -H "Content-Type: application/json" \
  -d '{"model":"claude-opus-5-5","max_tokens":16,"messages":[{"role":"user","content":"Reply with OK only."}]}' 
```

### If something goes wrong

**Key missing or rejected (401)**

Copy your current key from API keys. Paste it only into the app’s key field or your local key variable, then save and reopen the app.

**Not enough available credit (402)**

Check Balance and the key’s spending cap. Another request may have temporarily reserved credit; read the message before adding funds.

**Wrong URL, model or limit (400/404)**

Copy the exact values above. Do not add an operation such as /chat/completions to a Base URL. For GLM, context is 1000000 and maximum output is 65536.

**Busy or rate limited (429)**

Wait for the active request to finish or the rate limit to reset. Avoid repeatedly submitting the same request.

**Interrupted request or uncertain charge**

Check Activity first. Keep the request ID for support; never include your key in a message or screenshot.

**Model choices and service notes**

Claude Code uses the Anthropic Messages API on the bare host https://api.outrun.to (no /v1). All Claude models are set up: claude-opus-5-5 (default), claude-opus-5, claude-sonnet-5-5, claude-sonnet-5, claude-fable-5-1 and claude-fable-5. Each also has a lower-cost route 2: add -2 to the ID (for example claude-opus-5-5-2). If a request returns 402, check your account balance and key spending cap. For 429, read the error: wait for the active request to finish or the rate limit to reset. If a paid request is interrupted or its outcome is uncertain, check Activity before trying again; do not automatically repeat it. Choose an available model from Models & Pricing; paused models cannot be used.

[Contact @outrunsupport](https://t.me/outrunsupport)

============================================================
Source: https://outrun.to/docs/setup/codex (Codex CLI setup)
============================================================

# Outrun API Setup for Codex CLI

> Configure Codex CLI with Outrun using the documented provider settings, API base URL and model IDs. Includes platform-specific setup steps.

Canonical: https://outrun.to/docs/setup/codex



## Windows

Follow these steps in a terminal window. Choose your computer below.

### Before you start

Have your API key and some Outrun credit ready. Keep your key private.

[Get your API key](https://outrun.to/api-keys)

[Check balance](https://outrun.to/balance)

[Choose a model](https://outrun.to/docs/models)

Open PowerShell: open Start, type PowerShell, then open Windows PowerShell. Commands below are for this window.

[Download Desktop App](https://github.com/openai/codex/releases)

[GitHub Repository](https://github.com/openai/codex)

**Optional setup helper Use an AI assistant to walk through these steps**

Copy the setup instructions into an AI assistant you already use. Review any file changes it suggests before allowing them.

Enter your API key directly in the app or local settings. Keep it out of the assistant’s chat.

**Connection settings Show ⌄**

Connection settings

```
Base URL in config.toml: https://api.outrun.to/v1
Key variable: OUTRUN_API_KEY
First model: gpt-6.1-sol
```

Get your key from Outrun → API keys and enter it privately in the app or local settings. Test with “Reply with OK only.” Tests use credit; check Activity before retrying.

### Setup Steps

1. 1

  **Install Node.js first**

  Use the Download Node.js link below and choose the LTS version for your computer. Follow its installation instructions, then close and reopen your terminal. Node.js includes npm, the program used by the Install Command below.

  [Download Node.js (LTS)](https://nodejs.org/en/download)

2. 2

  **Install the tool**

  Open Commands and settings below. Copy only the Install Command for your selected computer into your terminal and press Enter. Wait for it to finish.

3. 3

  **Open the settings file**

  Find %USERPROFILE%\.codex\config.toml. Press Windows+R, paste the folder part of this path, and press Enter. %USERPROFILE% means your Windows user folder. Open the file in a text editor such as Notepad. If the folder or file is missing, create it.

4. 4

  **Add the Outrun settings**

  Open Commands and settings below and copy the Configuration Template into this file. If the file already has settings, keep them and add only the Outrun provider/settings block. Save the file. Do not paste JSON or TOML settings into the terminal.

5. 5

  **Enter your key privately**

  Use the Environment Setup commands for your computer below. Replace KEY with your own Outrun key in your local terminal or settings file. Never put the key in an AI chat or screenshot.

6. 6

  **Reopen and test**

  Open a new terminal, type codex, and press Enter. The first model is gpt-6.1-sol. Start a new chat and send: Reply with OK only. A short reply with no connection error means the first test worked. This request uses Outrun credit; check Activity if it is interrupted before trying again.

**Commands and settings for Windows Show ⌄**

Install Command

```
npm install -g @openai/codex
```

Windows PowerShell · OUTRUN_API_KEY

```
# Save the key as a user environment variable (replace KEY with your own key).
[Environment]::SetEnvironmentVariable("OUTRUN_API_KEY", "KEY", "User")
# Close and reopen PowerShell / the Codex app so the change takes effect.

# Check in a NEW PowerShell window (prints only the first characters):
$env:OUTRUN_API_KEY.Substring(0, 6)

# If Codex still reports a missing key, use the experimental_bearer_token
# alternative shown in the config template below.
```

Replace KEY with your own key only in your local app, editor or terminal. Never send that key to an AI chat. Copy one command block at a time.

### Configuration Template

Config TOML

```
# Codex config file
#   macOS / Linux: ~/.codex/config.toml
#   Windows:       %USERPROFILE%\.codex\config.toml
# Base URL ends in /v1. Do not append /responses.

model = "gpt-6.1-sol"
model_provider = "outrun"
model_reasoning_effort = "high"

[model_providers.outrun]
name = "Outrun"
base_url = "https://api.outrun.to/v1"
wire_api = "responses"
env_key = "OUTRUN_API_KEY"

# Windows alternative (only if Codex cannot see the OUTRUN_API_KEY variable):
# delete the env_key line above and use this line instead, with your key in place of KEY.
# experimental_bearer_token = "KEY"

# All Outrun GPT models (gpt-6.1-sol is the default above).
# Switch with: codex -p gpt-6-astra   (or: codex -m gpt-6-astra)
[profiles."gpt-5.6-sol"]
model = "gpt-5.6-sol"
model_provider = "outrun"

[profiles."gpt-6.1-sol"]
model = "gpt-6.1-sol"
model_provider = "outrun"

[profiles."gpt-6-sol"]
model = "gpt-6-sol"
model_provider = "outrun"

[profiles."gpt-6-astra"]
model = "gpt-6-astra"
model_provider = "outrun"

[profiles."gpt-5.6-terra"]
model = "gpt-5.6-terra"
model_provider = "outrun"
```

**Optional: test with a script (uses credit) Show**

Test cURL

```
curl https://api.outrun.to/v1/responses \
  -H "Authorization: Bearer $OUTRUN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"gpt-6.1-sol","input":"Reply with OK only.","max_output_tokens":16}' 
```

### If something goes wrong

**Key missing or rejected (401)**

Copy your current key from API keys. Paste it only into the app’s key field or your local key variable, then save and reopen the app.

**Not enough available credit (402)**

Check Balance and the key’s spending cap. Another request may have temporarily reserved credit; read the message before adding funds.

**Wrong URL, model or limit (400/404)**

Copy the exact values above. Do not add an operation such as /chat/completions to a Base URL. For GLM, context is 1000000 and maximum output is 65536.

**Busy or rate limited (429)**

Wait for the active request to finish or the rate limit to reset. Avoid repeatedly submitting the same request.

**Interrupted request or uncertain charge**

Check Activity first. Keep the request ID for support; never include your key in a message or screenshot.

**Model choices and service notes**

Codex talks to Outrun through a custom model provider. All GPT models are set up: gpt-6.1-sol (default), gpt-6-sol, gpt-6-astra, gpt-5.6-sol and gpt-5.6-terra. Switch with codex -p <model> or codex -m <model>. If a request returns 402, check your account balance and key spending cap. For 429, read the error: wait for the active request to finish or the rate limit to reset. If a paid request is interrupted or its outcome is uncertain, check Activity before trying again; do not automatically repeat it. Choose an available model from Models & Pricing; paused models cannot be used.

[Contact @outrunsupport](https://t.me/outrunsupport)

## macOS

Follow these steps in a terminal window. Choose your computer below.

### Before you start

Have your API key and some Outrun credit ready. Keep your key private.

[Get your API key](https://outrun.to/api-keys)

[Check balance](https://outrun.to/balance)

[Choose a model](https://outrun.to/docs/models)

Open Terminal: press Command+Space, type Terminal, then press Enter. Commands below are for this window.

[Download Desktop App](https://github.com/openai/codex/releases)

[GitHub Repository](https://github.com/openai/codex)

**Optional setup helper Use an AI assistant to walk through these steps**

Copy the setup instructions into an AI assistant you already use. Review any file changes it suggests before allowing them.

Enter your API key directly in the app or local settings. Keep it out of the assistant’s chat.

**Connection settings Show ⌄**

Connection settings

```
Base URL in config.toml: https://api.outrun.to/v1
Key variable: OUTRUN_API_KEY
First model: gpt-6.1-sol
```

Get your key from Outrun → API keys and enter it privately in the app or local settings. Test with “Reply with OK only.” Tests use credit; check Activity before retrying.

### Setup Steps

1. 1

  **Install Node.js first**

  Use the Download Node.js link below and choose the LTS version for your computer. Follow its installation instructions, then close and reopen your terminal. Node.js includes npm, the program used by the Install Command below.

  [Download Node.js (LTS)](https://nodejs.org/en/download)

2. 2

  **Install the tool**

  Open Commands and settings below. Copy only the Install Command for your selected computer into your terminal and press Enter. Wait for it to finish.

3. 3

  **Open the settings file**

  Find ~/.codex/config.toml. In Finder choose Go → Go to Folder and paste the folder part of this path. ~ means your home folder. Open the file in a plain-text editor. If the folder or file is missing, create it.

4. 4

  **Add the Outrun settings**

  Open Commands and settings below and copy the Configuration Template into this file. If the file already has settings, keep them and add only the Outrun provider/settings block. Save the file. Do not paste JSON or TOML settings into the terminal.

5. 5

  **Enter your key privately**

  Use the Environment Setup commands for your computer below. Replace KEY with your own Outrun key in your local terminal or settings file. Never put the key in an AI chat or screenshot.

6. 6

  **Reopen and test**

  Open a new terminal, type codex, and press Enter. The first model is gpt-6.1-sol. Start a new chat and send: Reply with OK only. A short reply with no connection error means the first test worked. This request uses Outrun credit; check Activity if it is interrupted before trying again.

**Commands and settings for macOS Show ⌄**

Install Command

```
npm install -g @openai/codex
```

macOS / Linux · OUTRUN_API_KEY

```
# Put your key in the OUTRUN_API_KEY variable (replace KEY with your own key).
# Add the line to whichever shell files you use, then open a new terminal.
echo 'export OUTRUN_API_KEY="KEY"' >> ~/.zshrc
echo 'export OUTRUN_API_KEY="KEY"' >> ~/.bashrc
source ~/.zshrc 2>/dev/null || source ~/.bashrc

# Check (prints only the first characters):
```

Replace KEY with your own key only in your local app, editor or terminal. Never send that key to an AI chat. Copy one command block at a time.

### Configuration Template

Config TOML

```
# Codex config file
#   macOS / Linux: ~/.codex/config.toml
#   Windows:       %USERPROFILE%\.codex\config.toml
# Base URL ends in /v1. Do not append /responses.

model = "gpt-6.1-sol"
model_provider = "outrun"
model_reasoning_effort = "high"

[model_providers.outrun]
name = "Outrun"
base_url = "https://api.outrun.to/v1"
wire_api = "responses"
env_key = "OUTRUN_API_KEY"

# Windows alternative (only if Codex cannot see the OUTRUN_API_KEY variable):
# delete the env_key line above and use this line instead, with your key in place of KEY.
# experimental_bearer_token = "KEY"

# All Outrun GPT models (gpt-6.1-sol is the default above).
# Switch with: codex -p gpt-6-astra   (or: codex -m gpt-6-astra)
[profiles."gpt-5.6-sol"]
model = "gpt-5.6-sol"
model_provider = "outrun"

[profiles."gpt-6.1-sol"]
model = "gpt-6.1-sol"
model_provider = "outrun"

[profiles."gpt-6-sol"]
model = "gpt-6-sol"
model_provider = "outrun"

[profiles."gpt-6-astra"]
model = "gpt-6-astra"
model_provider = "outrun"

[profiles."gpt-5.6-terra"]
model = "gpt-5.6-terra"
model_provider = "outrun"
```

**Optional: test with a script (uses credit) Show**

Test cURL

```
curl https://api.outrun.to/v1/responses \
  -H "Authorization: Bearer $OUTRUN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"gpt-6.1-sol","input":"Reply with OK only.","max_output_tokens":16}' 
```

### If something goes wrong

**Key missing or rejected (401)**

Copy your current key from API keys. Paste it only into the app’s key field or your local key variable, then save and reopen the app.

**Not enough available credit (402)**

Check Balance and the key’s spending cap. Another request may have temporarily reserved credit; read the message before adding funds.

**Wrong URL, model or limit (400/404)**

Copy the exact values above. Do not add an operation such as /chat/completions to a Base URL. For GLM, context is 1000000 and maximum output is 65536.

**Busy or rate limited (429)**

Wait for the active request to finish or the rate limit to reset. Avoid repeatedly submitting the same request.

**Interrupted request or uncertain charge**

Check Activity first. Keep the request ID for support; never include your key in a message or screenshot.

**Model choices and service notes**

Codex talks to Outrun through a custom model provider. All GPT models are set up: gpt-6.1-sol (default), gpt-6-sol, gpt-6-astra, gpt-5.6-sol and gpt-5.6-terra. Switch with codex -p <model> or codex -m <model>. If a request returns 402, check your account balance and key spending cap. For 429, read the error: wait for the active request to finish or the rate limit to reset. If a paid request is interrupted or its outcome is uncertain, check Activity before trying again; do not automatically repeat it. Choose an available model from Models & Pricing; paused models cannot be used.

[Contact @outrunsupport](https://t.me/outrunsupport)

## Linux

Follow these steps in a terminal window. Choose your computer below.

### Before you start

Have your API key and some Outrun credit ready. Keep your key private.

[Get your API key](https://outrun.to/api-keys)

[Check balance](https://outrun.to/balance)

[Choose a model](https://outrun.to/docs/models)

Open your computer’s Terminal app. Commands below are for this window.

[**Download Desktop App](https://github.com/openai/codex/releases)

[**GitHub Repository](https://github.com/openai/codex)

**Optional setup helper Use an AI assistant to walk through these steps**

Copy the setup instructions into an AI assistant you already use. Review any file changes it suggests before allowing them.

Enter your API key directly in the app or local settings. Keep it out of the assistant’s chat.

**Connection settings Show ⌄**

Connection settings

```
Base URL in config.toml: https://api.outrun.to/v1
Key variable: OUTRUN_API_KEY
First model: gpt-6.1-sol
```

Get your key from Outrun → API keys and enter it privately in the app or local settings. Test with “Reply with OK only.” Tests use credit; check Activity before retrying.

### Setup Steps

1. 1

  **Install Node.js first**

  Use the Download Node.js link below and choose the LTS version for your computer. Follow its installation instructions, then close and reopen your terminal. Node.js includes npm, the program used by the Install Command below.

  [**Download Node.js (LTS)](https://nodejs.org/en/download)

2. 2

  **Install the tool**

  Open Commands and settings below. Copy only the Install Command for your selected computer into your terminal and press Enter. Wait for it to finish.

3. 3

  **Open the settings file**

  Find ~/.codex/config.toml. Open your home folder in the file manager and show hidden files (often Ctrl+H). ~ means your home folder; folders starting with a dot are hidden. Open the file in a text editor. If the folder or file is missing, create it.

4. 4

  **Add the Outrun settings**

  Open Commands and settings below and copy the Configuration Template into this file. If the file already has settings, keep them and add only the Outrun provider/settings block. Save the file. Do not paste JSON or TOML settings into the terminal.

5. 5

  **Enter your key privately**

  Use the Environment Setup commands for your computer below. Replace KEY with your own Outrun key in your local terminal or settings file. Never put the key in an AI chat or screenshot.

6. 6

  **Reopen and test**

  Open a new terminal, type codex, and press Enter. The first model is gpt-6.1-sol. Start a new chat and send: Reply with OK only. A short reply with no connection error means the first test worked. This request uses Outrun credit; check Activity if it is interrupted before trying again.

**Commands and settings for Linux Show ⌄**

Install Command

```
npm install -g @openai/codex
```

macOS / Linux · OUTRUN_API_KEY

```
# Put your key in the OUTRUN_API_KEY variable (replace KEY with your own key).
# Add the line to whichever shell files you use, then open a new terminal.
echo 'export OUTRUN_API_KEY="KEY"' >> ~/.zshrc
echo 'export OUTRUN_API_KEY="KEY"' >> ~/.bashrc
source ~/.zshrc 2>/dev/null || source ~/.bashrc

# Check (prints only the first characters):
```

Replace KEY with your own key only in your local app, editor or terminal. Never send that key to an AI chat. Copy one command block at a time.

### Configuration Template

Config TOML

```
# Codex config file
#   macOS / Linux: ~/.codex/config.toml
#   Windows:       %USERPROFILE%\.codex\config.toml
# Base URL ends in /v1. Do not append /responses.

model = "gpt-6.1-sol"
model_provider = "outrun"
model_reasoning_effort = "high"

[model_providers.outrun]
name = "Outrun"
base_url = "https://api.outrun.to/v1"
wire_api = "responses"
env_key = "OUTRUN_API_KEY"

# Windows alternative (only if Codex cannot see the OUTRUN_API_KEY variable):
# delete the env_key line above and use this line instead, with your key in place of KEY.
# experimental_bearer_token = "KEY"

# All Outrun GPT models (gpt-6.1-sol is the default above).
# Switch with: codex -p gpt-6-astra   (or: codex -m gpt-6-astra)
[profiles."gpt-5.6-sol"]
model = "gpt-5.6-sol"
model_provider = "outrun"

[profiles."gpt-6.1-sol"]
model = "gpt-6.1-sol"
model_provider = "outrun"

[profiles."gpt-6-sol"]
model = "gpt-6-sol"
model_provider = "outrun"

[profiles."gpt-6-astra"]
model = "gpt-6-astra"
model_provider = "outrun"

[profiles."gpt-5.6-terra"]
model = "gpt-5.6-terra"
model_provider = "outrun"
```

**Optional: test with a script (uses credit) Show**

Test cURL

```
curl https://api.outrun.to/v1/responses \
  -H "Authorization: Bearer $OUTRUN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"gpt-6.1-sol","input":"Reply with OK only.","max_output_tokens":16}' 
```

### If something goes wrong

**Key missing or rejected (401)**

Copy your current key from API keys. Paste it only into the app’s key field or your local key variable, then save and reopen the app.

**Not enough available credit (402)**

Check Balance and the key’s spending cap. Another request may have temporarily reserved credit; read the message before adding funds.

**Wrong URL, model or limit (400/404)**

Copy the exact values above. Do not add an operation such as /chat/completions to a Base URL. For GLM, context is 1000000 and maximum output is 65536.

**Busy or rate limited (429)**

Wait for the active request to finish or the rate limit to reset. Avoid repeatedly submitting the same request.

**Interrupted request or uncertain charge**

Check Activity first. Keep the request ID for support; never include your key in a message or screenshot.

**Model choices and service notes**

Codex talks to Outrun through a custom model provider. All GPT models are set up: gpt-6.1-sol (default), gpt-6-sol, gpt-6-astra, gpt-5.6-sol and gpt-5.6-terra. Switch with codex -p <model> or codex -m <model>. If a request returns 402, check your account balance and key spending cap. For 429, read the error: wait for the active request to finish or the rate limit to reset. If a paid request is interrupted or its outcome is uncertain, check Activity before trying again; do not automatically repeat it. Choose an available model from Models & Pricing; paused models cannot be used.

[Contact @outrunsupport](https://t.me/outrunsupport)

============================================================
Price list
============================================================

All 100 model IDs that are live in the price feed dated 2026-10-29, USD, exact billed rates (the web pages show them rounded up to the cent). Live JSON: https://api.outrun.to/pricing

- `gpt-6-astra` (GPT-6 Astra): up to 272k tokens: $0.588 in / $2.94 out per 1M tokens, cached input $0.0588; over 272k tokens: $1.176 in / $4.41 out per 1M tokens, cached input $0.1176
- `gpt-5.6-sol` (GPT-5.6 Sol): up to 272k tokens: $0.294 in / $1.764 out per 1M tokens, cached input $0.0294; over 272k tokens: $0.588 in / $2.646 out per 1M tokens, cached input $0.0588
- `gpt-5.6-terra` (GPT-5.6 Terra): up to 272k tokens: $0.1176 in / $0.7056 out per 1M tokens, cached input $0.01176; over 272k tokens: $0.2352 in / $1.0584 out per 1M tokens, cached input $0.02352
- `gpt-6-sol` (GPT-6 Sol): up to 272k tokens: $0.1176 in / $0.588 out per 1M tokens, cached input $0.01176; over 272k tokens: $0.2352 in / $0.882 out per 1M tokens, cached input $0.02352
- `gpt-6.1-sol` (GPT-6.1 Sol): up to 272k tokens: $0.1176 in / $0.588 out per 1M tokens, cached input $0.00588; over 272k tokens: $0.2352 in / $0.882 out per 1M tokens, cached input $0.01176
- `claude-opus-5-5` (Claude Opus 5.5): $2.24 in / $11.2 out per 1M tokens, cached input $0.112
- `claude-opus-5` (Claude Opus 5): $3.15 in / $15.75 out per 1M tokens, cached input $0.315
- `claude-sonnet-5` (Claude Sonnet 5): $1.12 in / $5.6 out per 1M tokens, cached input $0.112
- `claude-fable-5-1` (Claude Fable 5.1): $6.3 in / $31.5 out per 1M tokens, cached input $0.1575
- `claude-fable-5` (Claude Fable 5): $5.6 in / $28 out per 1M tokens, cached input $0.56
- `claude-haiku-5-5` (Claude Haiku 5.5): up to 100k tokens: $0.084 in / $0.42 out per 1M tokens, cached input $0.0084; over 100k tokens: $0.42 in / $2.1 out per 1M tokens, cached input $0.042
- `gemini-3.8-flash` (Gemini 3.8 Flash): $0.315 in / $1.575 out per 1M tokens, cached input $0.0315
- `grok-4.7` (Grok 4.7): up to 200k tokens: $0.084 in / $0.252 out per 1M tokens, cached input $0.021; over 200k tokens: $0.168 in / $0.504 out per 1M tokens, cached input $0.042
- `gemini-3-pro-image` (Nano Banana Pro / Gemini 3 Pro Image): $0.036 per image
- `gemini-3.1-flash-image` (Nano Banana / Gemini 3.1 Flash Image): $0.027 per image
- `gpt-image-2` (GPT Image 2): $0.06 per image
- `grok-imagine-image` (Grok Imagine Image): $0.06 per image
- `gemini-3-stt` (Gemini 3 STT): $0.315 in / $1.89 out per 1M tokens, cached input $0.0315
- `gemini-3.1-tts` (Gemini 3.1 TTS): $0.036 per request
- `gemini-omni-flash` (Gemini Omni Flash): $0.06 per second
- `veo-3.1` (Veo 3.1): 720p/1080p: $0.075 per second; 4K: $0.1125 per second
- `veo-3.1-fast` (Veo 3.1 Fast): 720p/1080p: $0.03 per second; 4K: $0.07 per second
- `seedance-2.5-720p` (Seedance 2.5 720p): $0.18 per second
- `seedance-2.5-1080p` (Seedance 2.5 1080p): $0.464 per second
- `seedance-2.0-720p` (Seedance 2.0 720p): $0.12474 per second
- `seedance-2.0-1080p` (Seedance 2.0 1080p): $0.31185 per second
- `grok-imagine-video-1.5-1080p` (Grok Imagine Video 1.5 1080p): $0.09 per second
- `wan3.0-video-1080p` (Wan 3.0 1080p): $0.33 per second
- `grok-imagine-video` (Grok Imagine Video): $0.03 per second
- `deepseek-v4-pro` (DeepSeek V4 Pro): $1.35 in / $4.05 out per 1M tokens, cached input $0.045
- `deepseek-v4-flash` (DeepSeek V4 Flash): $0.45 in / $1.35 out per 1M tokens, cached input $0.015
- `deepseek-v4.1-flash` (DeepSeek V4.1 Flash): $0.3 in / $1.2 out per 1M tokens, cached input $0.006
- `glm-5.3-flash` (GLM-5.3 Flash): $0.12 in / $0.42 out per 1M tokens, cached input $0.0345
- `claude-sonnet-5-5` (Claude Sonnet 5.5): $0.49 in / $2.45 out per 1M tokens, cached input $0.0245
- `glm-5.3-flash-uncensored` (GLM-5.3 Flash Uncensored): $0.05 in / $0.05 out per 1M tokens, cached input $0.005
- `glm-5.3-uncensored` (GLM-5.3 Uncensored): $0.1 in / $0.1 out per 1M tokens, cached input $0.02
- `deepseek-v4-pro-uncensored` (DeepSeek V4 Pro Uncensored): $0.5 in / $0.5 out per 1M tokens, cached input $0.1
- `glm-5.3-abliterated` (GLM-5.3 Abliterated): $1.4 in / $1.4 out per 1M tokens, cached input $0.4
- `deepseek-v4.1-flash-abliterated-fast` (DeepSeek V4.1 Flash Abliterated Fast): $1.4 in / $1.4 out per 1M tokens, cached input $0.32
- `mimo-v2.6-flash-uncensored-thinking` (MiMo V2.6 Flash Uncensored Thinking): $1 in / $3 out per 1M tokens, cached input $0.399
- `mimo-v2.6-flash-abliterated` (MiMo V2.6 Flash Abliterated): $1 in / $3 out per 1M tokens, cached input $0.55
- `mimo-v2.6-flash-uncensored` (MiMo V2.6 Flash Uncensored): $1 in / $3 out per 1M tokens, cached input $0.399
- `kimi-k3` (Kimi K3): $1.8 in / $9 out per 1M tokens, cached input $0.18
- `glm-5.3` (GLM-5.3): $0.6552 in / $2.0592 out per 1M tokens, cached input $0.1216
- `MiniMax-M3` (MiniMax M3): $0.18 in / $0.72 out per 1M tokens, cached input $0.036
- `glm-5.3-uncensored-2` (GLM-5.3 Uncensored (2)): $0.3492 in / $0.3492 out per 1M tokens, cached input $0.099
- `glm-5.3-flash-uncensored-2` (GLM-5.3 Flash Uncensored (2)): $0.168 in / $0.168 out per 1M tokens, cached input $0.0168
- `glm-5.3-abliterated-2` (GLM-5.3 Abliterated (2)): $4.5 in / $4.5 out per 1M tokens, cached input $1.32
- `glm-5.3-flash-abliterated` (GLM-5.3 Flash Abliterated): $1.32 in / $1.32 out per 1M tokens, cached input $0.36
- `deepseek-v4.1-flash-1017-2` (DeepSeek V4.1 Flash 1017 (2)): $0.388 in / $2.328 out per 1M tokens, cached input $0.0116
- `deepseek-v4.1-flash-abliterated` (DeepSeek V4.1 Flash Abliterated): $2.1 in / $2.1 out per 1M tokens, cached input $0.36
- `glm-5.3-flash-code-uncensored` (GLM-5.3 Flash Code Uncensored): $12 in / $12 out per 1M tokens, cached input $12
- `gpt-oss-120b-uncensored` (GPT OSS 120B Uncensored): $0.219 in / $1.456 out per 1M tokens, cached input $0.219
- `minimax-m3-uncensored` (MiniMax M3 Uncensored): $1.5 in / $8 out per 1M tokens, cached input $0.375
- `outrun-redteam-flash` (Outrun Red Team Flash): $1.35 in / $4.5 out per 1M tokens, cached input $0.23145
- `outrun-redteam-pro` (Outrun Red Team Pro): $2.25 in / $9 out per 1M tokens, cached input $0.6375
- `hermes-4-405b-uncensored` (Hermes 4 405B): $1.8 in / $7.2 out per 1M tokens, cached input $1.8
- `nemotron-3-ultra-550b-a55b-uncensored` (Nemotron 3 Ultra 550B Uncensored): $1.6875 in / $9.9 out per 1M tokens, cached input $0.423
- `glm-5.3-uncensored-3` (GLM-5.3 Uncensored (3)): $2.5 in / $4.5 out per 1M tokens, cached input $0.6
- `glm-5.3-flash-uncensored-3` (GLM-5.3 Flash Uncensored (3)): $0.4 in / $1.6 out per 1M tokens, cached input $0.14
- `glm-5.3-flash-cybersecurity` (GLM-5.3 Flash Cybersecurity): $0.3 in / $1 out per 1M tokens, cached input $0.15
- `deepseek-v4.1-flash-1017-3` (DeepSeek V4.1 Flash 1017 (3)): $0.2 in / $1.2 out per 1M tokens, cached input $0.01
- `deepseek-v4.1-flash-1017-thinking-3` (DeepSeek V4.1 Flash 1017 Thinking (3)): $0.2 in / $1.2 out per 1M tokens, cached input $0.01
- `deepseek-v4.1-flash-1017-uncensored-3` (DeepSeek V4.1 Flash Uncensored 1017 · Route 3): $0.66 in / $2.64 out per 1M tokens, cached input $0.0165
- `deepseek-v4.1-flash-1017-uncensored-thinking-3` (DeepSeek V4.1 Flash Uncensored Thinking 1017 · Route 3): $0.66 in / $2.64 out per 1M tokens, cached input $0.0165
- `glm-5.3-flash-apex-abliterated` (Apex Flash 1 Abliterated): $0.3 in / $1 out per 1M tokens, cached input $0.15
- `anime-image-uncensored` (Anime Image Uncensored): $0.06 per image
- `chroma-uncensored` (Chroma Uncensored): $0.16 per image
- `lustify-sdxl-uncensored` (Lustify SDXL Uncensored): $0.0626 per image
- `seedream-5.0-lite-edit-uncensored` (Seedream 5.0 Lite Edit Uncensored): $0.0626 per image
- `vision-image-uncensored` (Vision Image Uncensored): $0.08 per image
- `vision-pro-image-uncensored` (Vision Pro Image Uncensored): $0.1876 per image
- `cyberrealistic-xl-uncensored` (CyberRealistic XL Uncensored): 2048x2048: $0.0544 per image; 512x512: $0.0272 per image; other sizes: $0.0408 per image
- `illustrious-xl-uncensored` (WAI Illustrious SDXL Uncensored): 2048x2048: $0.0544 per image; 512x512: $0.0272 per image; other sizes: $0.0408 per image
- `qwen-image-2.1-uncensored` (Qwen Image 2.1 Uncensored): 1K: $0.075 per image; 2K: $0.4 per image
- `MiniMax-H3` (MiniMax H3): $0.06 per second
- `wan-3.0-video-uncensored` (Wan 3.0 Video Uncensored): 480p: $0.4 per second; 720p: $0.8 per second; 1080p: $1.6 per second
- `mercury-3` (Mercury 3): $0.2 in / $0.7 out per 1M tokens, cached input $0.02
- `gemini-4-argon` (Gemini 4 Argon): $1.2 in / $6 out per 1M tokens, cached input $0.12
- `deepseek-v4.1-flash-abliterated-2` (DeepSeek V4.1 Flash Abliterated (2)): $0.3 in / $0.3 out per 1M tokens, cached input $0.1
- `minimax-m3.1-flash` (MiniMax M3.1 Flash): $0.09 in / $0.36 out per 1M tokens, cached input $0.009
- `gemini-3.7-flash` (Gemini 3.7 Flash): $0.945 in / $4.725 out per 1M tokens, cached input $0.0945
- `gemini-3.1-pro-preview` (Gemini 3.1 Pro Preview): up to 200k tokens: $1.26 in / $7.56 out per 1M tokens, cached input $0.126; over 200k tokens: $2.52 in / $11.34 out per 1M tokens, cached input $0.252
- `kimi-k3.1-pro` (Kimi K3.1 Pro): $0.2 in / $0.2 out per 1M tokens, cached input $0.06
- `qwen3.7-max` (Qwen 3.7 Max): with cache_control: $1.44 in / $4.32 out per 1M tokens, cached input $0.144; standard: $1.44 in / $4.32 out per 1M tokens, cached input $0.288
- `qwen3.8-flash` (Qwen 3.8 Flash): with cache_control: $0.096 in / $0.324 out per 1M tokens, cached input $0.012; standard: $0.096 in / $0.324 out per 1M tokens, cached input $0.012
- `qwen3.8-max` (Qwen 3.8 Max): with cache_control: $1.44 in / $4.32 out per 1M tokens, cached input $0.18; standard: $1.44 in / $4.32 out per 1M tokens, cached input $0.18
- `qwen3.7-plus` (Qwen 3.7 Plus): with cache_control, up to 256k tokens: $0.285 in / $1.14 out per 1M tokens, cached input $0.0285; with cache_control, over 256k tokens: $0.855 in / $3.42 out per 1M tokens, cached input $0.0855; up to 256k tokens: $0.285 in / $1.14 out per 1M tokens, cached input $0.057; over 256k tokens: $0.855 in / $3.42 out per 1M tokens, cached input $0.171
- `gpt-image-2.5-flare` (GPT Image 2.5 Flare): $0.015 per image
- `gpt-image-2.5-sunburst` (GPT Image 2.5 Sunburst): $0.015 per image
- `doubao-seedream-5-0-pro` (Seedream 5 Pro): 1K: $0.04275 per image; 2K: $0.0855 per image
- `claude-opus-5-5-2` (Claude Opus 5.5 (2)): $0.8 in / $4 out per 1M tokens, cached input $0.04
- `claude-opus-5-2` (Claude Opus 5 (2)): $1 in / $5 out per 1M tokens, cached input $0.1
- `claude-sonnet-5-5-2` (Claude Sonnet 5.5 (2)): $0.4 in / $2 out per 1M tokens, cached input $0.02
- `claude-sonnet-5-2` (Claude Sonnet 5 (2)): $0.4 in / $2 out per 1M tokens, cached input $0.04
- `claude-fable-5-1-2` (Claude Fable 5.1 (2)): $2 in / $10 out per 1M tokens, cached input $0.05
- `claude-fable-5-2` (Claude Fable 5 (2)): $2 in / $10 out per 1M tokens, cached input $0.2
- `claude-haiku-5-5-2` (Claude Haiku 5.5 (2)): $0.02 in / $0.1 out per 1M tokens, cached input $0.002