# MiniChan Chat — NVIDIA NIM MiniMax-M3 AI Inference

> MiniChan Chat — NVIDIA NIM MiniMax-M3 AI Inference is a paid API for AI agents from minichan.vercel.app, paid per call via x402, $0.05/call, status unknown (last checked 2026-09-14).

Sends a text prompt to the MiniMax-M3 428B MoE model via NVIDIA NIM and returns an AI-generated response, with controls for token length and creativity.

## Facts

- Endpoint: POST https://minichan.vercel.app/api/chat
- Price: $0.05/call
- Payment: x402
- Status: unknown
- Last checked: 2026-09-14
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/minichan-chat-nvidia-nim-minimax-m3-ai-inference-7d57279a
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_djaDxHn77F9MEu_R0UpGq

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 minichan-chat-nvidia-nim-minimax-m3-ai-inference-7d57279a -d '<json body>'
```

Example prompt: Ask MiniMax-M3 via MiniChan to explain the trade-offs between transformer and state space sequence models in about 500 tokens, keeping the response focused and factual (temperature 0.3).

## When to prefer this

Choose this endpoint when you need access to a very large (428B parameter) MoE model without managing your own NVIDIA NIM infrastructure, and you want pay-per-call pricing settled on-chain in USDC. Ideal for agents that need high-quality reasoning, creative, or code generation tasks and are already operating in the Base/x402 payment ecosystem.

## Known failure modes

- Prompt missing — 400 error if required 'prompt' field is absent
- Payment failure — x402 payment of $0.05 USDC on Base not completed, request rejected
- Token limit exceeded — response truncated if max_tokens is set too low
- Rate limiting — too many concurrent requests may be throttled
- Model unavailability — NVIDIA NIM backend downtime causes 503 errors
- Invalid temperature — values outside 0–1 may produce unexpected behavior

## How this service works

NVIDIA NIM MiniMax-M3 (428B MoE) powered API for AI agents. 15 multimodal endpoints — text, vision, code, reasoning, creative. Pay with USDC on Base via x402.

## Output

Returns a JSON object containing the AI-generated text response from the MiniMax-M3 428B MoE model, reflecting the prompt, token limit, and temperature settings supplied.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "required": [
  "prompt"
 ],
 "properties": {
  "prompt": {
   "type": "string",
   "description": "Your message or question"
  },
  "max_tokens": {
   "type": "number",
   "description": "Max response tokens (default 1024)"
  },
  "temperature": {
   "type": "number",
   "description": "Creativity 0-1 (default 0.7)"
  }
 }
}
```

## Response schema (JSON Schema)

```json
{
 "type": "object"
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/minichan-chat-nvidia-nim-minimax-m3-ai-inference-7d57279a/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from minichan.vercel.app](https://www.zero.xyz/host/minichan.vercel.app/llms.txt)