# x402engine Kimi K3 (MoonshotAI)

> x402engine Kimi K3 (MoonshotAI) is a paid API for AI agents from x402engine.app, paid per call via x402, $0.04/call, status unknown (last checked 2026-10-02).

Runs inference on MoonshotAI's Kimi K3 frontier multimodal model with 1M token context window, optimized for agentic tasks, via pay-per-call micropayment

## Facts

- Endpoint: POST https://x402engine.app/api/llm/kimi-k3?utm_source=zero.xyz
- Price: $0.04/call
- Payment: x402
- Status: unknown
- Last checked: 2026-10-02
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/x402engine-kimi-k3-moonshotai-1577f54c
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_AHyjqaVQZswh-coFargEC

Status and success rate cover calls made through Zero and Zero's own probes. Third-party monitors may report differently.

## How to call it through Zero

Zero handles the 402 payment challenge and records the run. With the Zero CLI installed (`npm i -g @zeroxyz/cli`):

```sh
zero fetch --capability x402engine-kimi-k3-moonshotai-1577f54c -d '<json body>'
```

Example prompt: Use the Kimi K3 model to analyze this 500-page contract and identify all liability clauses — it needs a model with at least a 1M token context window and strong agentic reasoning.

## When to prefer this

Choose this endpoint when you need MoonshotAI's Kimi K3 specifically — especially for tasks requiring a very long context window (up to 1M tokens), multimodal inputs, or strong agentic task performance. Prefer it over other LLMs on x402engine when the task involves extremely long documents, large codebases, or complex multi-step reasoning that exceeds typical context limits. Ideal for agents that need frontier-quality reasoning without a direct subscription to MoonshotAI.

## Known failure modes

- Payment failure if insufficient USDC balance (HTTP 402)
- Context window exceeded if input exceeds 1M tokens
- Model timeout for extremely complex or long agentic tasks
- Malformed request body returns 400 error
- Rate limiting under high concurrent load

## How this service works

MoonshotAI's frontier multimodal model with 1M context for agentic tasks

## Output

Returns generated text output from MoonshotAI's Kimi K3 model, including completions, reasoning traces, agentic task outputs, or multimodal analysis results, depending on the prompt and input provided.

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/x402engine-kimi-k3-moonshotai-1577f54c/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from x402engine.app](https://www.zero.xyz/host/x402engine.app/llms.txt)