# AgentRXI IBM Granite 4.0 Micro Chat Completion

> AgentRXI IBM Granite 4.0 Micro Chat Completion is a paid API for AI agents from agentrxi.com, paid per call via x402, $0.002346/call, status unknown (last checked 2026-10-15).

Pay-per-call LLM chat completion using IBM Granite 4.0 Micro, returning OpenAI-compatible responses with no API key required

## Facts

- Endpoint: GET https://agentrxi.com/v1/granite/chat/completions?utm_source=zero.xyz
- Price: $0.002346/call
- Payment: x402
- Status: unknown
- Last checked: 2026-10-15
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/agentrxi-ibm-granite-4-0-micro-chat-completion-ce5e937f
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_jzGauluG7PVl5JM1AGwgd

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 agentrxi-ibm-granite-4-0-micro-chat-completion-ce5e937f
```

Example prompt: Use Granite 4.0 Micro to answer this question with a system prompt of 'You are a concise assistant' and a max of 200 tokens: 'What are three benefits of containerized deployments?'

## When to prefer this

Choose this endpoint when you need a lightweight, fast, and cheap LLM call without managing API keys or subscriptions — ideal for simple reasoning steps, translation, classification, or text generation in an agent pipeline. Prefer it over larger Granite or Llama 70B endpoints when the task is straightforward and latency or cost is a concern. Best for agents already using x402 micropayment flows on AgentRXI.

## Known failure modes

- Missing required 'prompt' query parameter returns a 400 error
- Payment not attached or insufficient USDC balance causes a 402 Payment Required response
- Temperature or max_tokens out of acceptable range may cause a 400 validation error
- Model overload or upstream IBM service outage returns a 503
- Prompt too long for model context window triggers a 400 or truncation

## How this service works

OpenAI-compatible chat completions on IBM Granite 4.0 Micro, the lowest-cost chat here, with tool calling (131K context). Set your OpenAI SDK base URL to https://agentrxi.com/v1/granite and pay per token with x402 instead of an API key: $0.051 per million input and $0.336 per million output tokens, minimum $0.002 a call, at most $0.0049. The 402 quotes your request's maximum; "upto" payers are charged only the tokens used. Same request/response shape as OpenAI (tool calling, 

## Output

An OpenAI-compatible JSON chat completion object containing the assistant's reply text, finish reason (e.g. 'stop'), model name ('granite-4.0-micro'), and token usage breakdown (prompt, completion, and total tokens).

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "$schema": "https://json-schema.org/draft/2020-12/schema",
 "required": [
  "input"
 ],
 "properties": {
  "input": {
   "type": "object",
   "required": [
    "type",
    "method"
   ],
   "properties": {
    "type": {
     "type": "string",
     "const": "http"
    },
    "method": {
     "enum": [
      "GET"
     ],
     "type": "string"
    },
    "queryParams": {
     "type": "object",
     "required": [
      "prompt"
     ],
     "properties": {
      "prompt": {
       "type": "string",
       "description": "User message"
      },
      "system": {
       "type": "string",
       "description": "Optional system message"
      },
      "max_tokens": {
       "type": "integer"
      },
      "temperature": {
       "type": "number"
      }
     }
    }
   },
   "additionalProperties": false
  },
  "output": {
   "type": "object",
   "required": [
    "type"
   ],
   "properties": {
    "type": {
     "type": "string"
    },
    "example": {
     "type": "object"
    }
   }
  }
 }
}
```

## Response schema (JSON Schema)

```json
{
 "type": "json",
 "example": {
  "id": "chatcmpl-...",
  "model": "granite-4.0-micro",
  "usage": {
   "total_tokens": 18,
   "prompt_tokens": 14,
   "completion_tokens": 4
  },
  "object": "chat.completion",
  "choices": [
   {
    "index": 0,
    "message": {
     "role": "assistant",
     "content": "Bonjour !"
    },
    "finish_reason": "stop"
   }
  ]
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/agentrxi-ibm-granite-4-0-micro-chat-completion-ce5e937f/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from agentrxi.com](https://www.zero.xyz/host/agentrxi.com/llms.txt)