# DeepSeek V3.2 Speciale Enhanced Reasoning LLM

> DeepSeek V3.2 Speciale Enhanced Reasoning LLM is a paid API for AI agents from x402-gateway-production.up.railway.app, paid per call via x402, $0.008000/call, status unknown (last checked 2026-09-15).

Runs DeepSeek's enhanced chain-of-thought reasoning variant (deepseek-v3.2-speciale) for complex multi-step tasks via a pay-per-call gateway

## Facts

- Endpoint: POST https://x402-gateway-production.up.railway.app/api/llm/deepseek-v3.2-speciale
- Price: $0.008000/call
- Payment: x402
- Status: unknown
- Last checked: 2026-09-15
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/x402-gateway-production-up-railway-app-8818f99e
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_ye6ZWC2L2VK7GaX_69eQi

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 x402-gateway-production-up-railway-app-8818f99e -d '<json body>'
```

Example prompt: Use DeepSeek's enhanced chain-of-thought reasoning model to solve this step by step: A train leaves Chicago at 60mph and another leaves New York at 80mph toward each other 1,000 miles apart — when and where do they meet?

## When to prefer this

Choose this endpoint when you need DeepSeek's enhanced reasoning model specifically — it includes visible reasoning_tokens in the response, making it ideal for complex multi-step math, logic, coding, or analysis tasks where chain-of-thought quality matters. Prefer over generic GPT-4 or Claude endpoints when you want DeepSeek's specific reasoning style at a low per-call cost ($0.008 USDC) without managing your own API keys.

## Known failure modes

- Invalid or missing messages array returns 400 error
- Payment not included or insufficient USDC causes x402 payment required error
- Upstream model unavailable returns 503 or timeout
- Malformed JSON body returns parse error
- Rate limiting if too many concurrent requests

## How this service works

DeepSeek's enhanced reasoning variant — improved chain-of-thought for complex tasks

## Output

Returns a text completion from deepseek-v3.2-speciale-20251201 with the model's response, token usage breakdown (prompt tokens, completion tokens, reasoning tokens), and upstream inference cost details in USDC.

## Example request

```json
{
 "messages": [
  {
   "role": "user",
   "content": "Solve this multi-step math problem step by step: If a store sells apples at $2 per pound and oranges at $3 per pound, and a customer buys 5 pounds of apples and 3 pounds of oranges, what is the total cost?"
  }
 ]
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/x402-gateway-production-up-railway-app-8818f99e/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from x402-gateway-production.up.railway.app](https://www.zero.xyz/host/x402-gateway-production.up.railway.app/llms.txt)