# GPT-5.6 Terra Chat Completions

> GPT-5.6 Terra Chat Completions is a paid API for AI agents from api.monpham.work, paid per call via x402, $0.025/call, status unknown (last checked 2026-10-27).

Runs chat completions via a balanced frontier LLM (GPT-5.6 Terra) optimized for everyday coding, multi-step reasoning, and agentic tasks, with up to 16K input and 4K output tokens.

## Facts

- Endpoint: GET https://api.monpham.work/v1/terra/chat/completions?utm_source=zero.xyz
- Price: $0.025/call
- Payment: x402
- Status: unknown
- Last checked: 2026-10-27
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/gpt-5-6-terra-chat-completions-25467f37
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_tDTg9_MfsXHqk7ihuHUe0

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 gpt-5-6-terra-chat-completions-25467f37
```

Example prompt: Use the Terra model to help me debug this Python script that parses nested JSON and write a corrected version with proper error handling — I need solid reasoning, not just the fastest response.

## When to prefer this

Choose Terra when you need more reasoning capability and accuracy than a fast/cheap model (like Luna) but want to avoid the highest cost of a top-tier model (like Sol). It's the right balance for everyday coding tasks, structured multi-step reasoning, moderately complex agentic pipelines, and technical Q&A where quality matters. Prefer it over lighter models when GPQA-level accuracy (92.5%) is needed, and over heavier models when cost is a concern.

## Known failure modes

- Input exceeds 16K token limit — request rejected or truncated
- Output hits 4K token limit mid-response — completion cut off
- Payment of $0.025 USDC not processed — 402 Payment Required error
- Malformed request body not matching OpenAI-compatible format — 400 Bad Request
- Model temporarily unavailable — 503 Service Unavailable
- Rate limit exceeded — 429 Too Many Requests

## How this service works

GPT-5.6 Terra — balanced frontier model (Intelligence Index 55.0, GPQA Diamond 92.5%, Terminal-Bench Hard 57.6%). Best for everyday coding, multi-step reasoning, and agentic tasks needing more capability than Luna without Sol's cost. Limits: 16K input / 4K output tokens per call. OpenAI-compatible request and response format.

## Output

An OpenAI-compatible chat completion response containing the model's generated text, finish reason, token usage statistics, and model metadata. The response follows the standard OpenAI ChatCompletion object format, including choices array with message content up to 4K output tokens.

## Request schema (JSON Schema)

```json
{
 "properties": {
  "input": {
   "type": "object",
   "required": [
    "type",
    "method"
   ],
   "properties": {
    "type": {
     "type": "string",
     "const": "http"
    },
    "method": {
     "enum": [
      "GET"
     ],
     "type": "string"
    }
   },
   "additionalProperties": false
  }
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/gpt-5-6-terra-chat-completions-25467f37/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from api.monpham.work](https://www.zero.xyz/host/api.monpham.work/llms.txt)