Call any frontier model through one endpoint
One OpenAI- and Anthropic-compatible endpoint for every frontier model. Provider list price on tokens, no markup, no subscription, priced before the call runs.
One LLM API for every frontier model — OpenAI, Anthropic, Google, DeepSeek, xAI
Send the JSON an OpenAI chat-completions client already sends — model, messages, stream, tools — and pick any model on the rate card by changing the model string. Anthropic-shaped requests go to the Messages-compatible path, and OpenAI's Responses shape has its own endpoint, so an existing client keeps its wire format.
Streaming, tool calls, vision inputs and multi-turn history behave as they do upstream; the response carries the usage the request actually consumed. Every model routes through the same endpoint, so switching models is a one-line change rather than a second integration.
How to call the LLM API: OpenAI-compatible chat completions, Anthropic-compatible messages
One endpoint, priced in dollars before it runs and settled in USDC. No API key and no account — the request pays for itself.
POST /api/v1/chat/completions
Host: blockrun.aiLLM API pricing per model, at the provider's own token rate
LLM API for AI agents: Claude Code, OpenClaw, Cursor, or your own code
In Claude Code, Claude Desktop, Cursor, Windsurf, Cline and any other MCP host, this is the blockrun_chat tool of the BlockRun MCP server — the agent calls it like any other tool and the result comes back in the conversation. In OpenClaw, Codex and Hermes, ClawRouter exposes the same catalog. In your own code, the Python and TypeScript SDKs wrap the endpoint above with types; any HTTP client works too. Every harness, with its setup →
No account and no API key stand between an agent and the first call: the request is quoted in dollars and the agent's client pays it. A team that prefers keys gets one, with a monthly invoice.
LLM API vs LiteLLM, Portkey, Helicone and the provider consoles
A developer who wants more than one model today integrates the providers one by one, or puts a router in front of them — LiteLLM, Portkey, Helicone and their peers. Those routers are good software, and every one of them still needs a key per provider behind it, an account with each vendor, and a bill from each. This endpoint is the router and the billing in one: one base URL, one payment, every model.
The other difference is who can call it. A key-based router needs a human to sign up, paste keys and top up a card; here a request pays for itself, so an agent with a funded balance can call any model on the card with no account on file. That is why the pricing is the provider's own token rate rather than a markup: the margin is not on tokens, it is a flat per-call fee.
What this is not: a proxy that hides which model answered. The response carries the model that served it and the usage it consumed, the same fields the provider returns, so a switch from one vendor to another is a diff in your logs, not a surprise on your invoice.
Related tools
LLM API questions: keys, compatibility, cost, free tier
Is this LLM API OpenAI-compatible?
Yes. Point an existing OpenAI SDK at the base URL and the first chat-completions call works unchanged — model, messages, stream, tools, response_format. Anthropic's Messages shape and OpenAI's Responses shape each have their own endpoint.
Which models can I call through one LLM API?
Every model on the rate card above, and the rest of the catalog: OpenAI, Anthropic, Google, DeepSeek, xAI, Qwen, Moonshot and more, all through the same endpoint by changing the model string.
How much does the LLM API cost per token?
The provider's published input and output rate per million tokens, with no markup on tokens, plus a flat per-call fee. Every request is quoted in dollars before it runs, so the cost is known before the tokens are.
Is there a free LLM API tier?
Yes. Several open-weight models are free and need no wallet or payment header at all — the free rows on the rate card are callable today, with per-IP rate limits instead of a bill.
Does the LLM API support streaming, tool calling and vision?
Yes. Server-sent streaming, function and tool calls, image inputs and multi-turn history behave as they do upstream, and the response carries the usage the request actually consumed.
LLM API vs LiteLLM, Portkey or OpenRouter — what is the difference?
A router still needs a key and an account per provider behind it, or a prepaid credit balance with the router. Here the request pays for itself at the provider's token rate, so there is nothing to sign up for and no credit to pre-load.
Can I switch models without changing my code?
Yes. Every model routes through the same endpoint, so switching from one vendor's model to another is a one-line change to the model string — no second SDK, no second integration.
Do I need an API key to use the LLM API?
No. Payment is the authentication — a request carries an x402 payment header and settles in USDC on Base or Solana. There is no signup, no dashboard and no key to rotate.
Can an AI agent call the LLM API without a human account?
Yes — that is what it is for. An agent with a funded balance pays per request from its own wallet; nothing has to be provisioned by a person first.
Two ways to run this
Get an API key
Sign in, get a key, and call the same endpoints with usage billed to your account. For teams that want one invoice instead of one wallet.
Get an API keyOr let your agent pay per call, in USDC
No account, no API key. Three steps, and the last one is the request this page is about.
- Point your agent at BlockRun
Install ClawRouter, or add the BlockRun MCP server to Claude Code, Cursor, Codex or OpenClaw. Any OpenAI-compatible client works too — change the base URL, nothing else.
curl -fsSL https://blockrun.ai/ClawRouter-update | bash - Fund a wallet with USDC
Send USDC to a wallet on Base or Solana. There is no account to create and no API key to issue: the wallet is the account, and it pays per call.
- Call LLM API
The first call returns HTTP 402 with the exact price, your agent signs it, and the answer comes back. Payment settles on-chain; nothing is charged if the call fails.
POST /api/v1/chat/completions Host: blockrun.ai
Explore everything on BlockRun
- GPT-6 Astra
- GPT-6.1 Sol
- GPT-6 Sol
- GPT-6 Luna
- GPT-5.6 Sol
- GPT-5.6 Terra
- GPT-5.6 Luna
- GPT-5.6 Sol Pro
- GPT-5.6 Terra Pro
- GPT-5.6 Luna Pro
- GPT-5.5
- GPT-5.5 Pro
- ChatGPT Instant (GPT-5.5)
- GPT-5.4
- GPT-5.4 Pro
- GPT-5.1
- GPT-5.2
- GPT-5.4 Mini
- GPT-5 Mini
- GPT-5.4 Nano
- GPT-5.2 Pro
- GPT-5.3 Codex
- GPT-4.1
- GPT-4.1 Mini
- GPT-4.1 Nano
- GPT-4o
- GPT-4o Mini
- o1
- o3
- o3-mini
- All models →