Name: OpenAI GPT-5 Pro Reasoning Chat Model API
Brand: magicapi
SKU: api-magicapi-ai-language-model-openai-gpt-5-pro
Availability: InStock
Author: API Market

Shashank Agarwal

# 🧠 Reasoning Chat Model API (OpenAI GPT-5 Pro)

![Reasoning Chat Model API (OpenAI GPT-5 Pro) on api.market](https://blog.api.market/wp-content/uploads/2026/10/openai__gpt-5-pro-cover-api-market.jpg)

## Overview
GPT-5 Pro is a language model from OpenAI with built-in thinking. Its developer describes it as putting expert-level intelligence in everyone's hands.

Send either a `prompt` (with an optional `system_prompt`) or a full `messages` list, plus optional image URLs in `image_input`. `verbosity` sets how long the answer is and `max_completion_tokens` caps the tokens generated. The model returns text as a list of chunks; join them for the full response.

## Features
- **Prompt or messages**: send a single `prompt`, or a `messages` list of `{"role", "content"}` objects for multi-turn input. When `messages` is set, `prompt` and `system_prompt` are ignored.
- **System prompt**: `system_prompt` sets the assistant's behavior for a `prompt` request.
- **Image input**: pass one or more image URLs in `image_input`.
- **Verbosity**: `low`, `medium` (default) or `high` answer length.
- **Output cap**: `max_completion_tokens` limits the tokens generated, including the tokens the model spends on reasoning. Leave it unset or set it high enough to avoid empty responses.

## Use Cases
1. **Explanations of hard topics**: step-by-step answers to math, science and engineering questions, as in the example.
2. **Analysis and planning**: work through problems that need several reasoning steps before an answer.
3. **Writing and editing**: drafts, rewrites and summaries with the length set by `verbosity`.
4. **Questions about images**: send screenshots, charts or photos in `image_input` and ask about them.

## Visual Examples
### Example Input
A one-line question with `verbosity` set to `medium`, no images and no messages.

```json
{
  "prompt": "Can you explain the Banach–Tarski paradox?",
  "messages": [],
  "verbosity": "medium",
  "image_input": []
}
```

### Example Output
Joined output (excerpt; the full answer is 3,733 characters):

> Here’s the idea, in layers.
>
> What the theorem says
> - In 3-dimensional space, you can cut a solid ball into finitely many disjoint pieces and, using only rotations and translations (no stretching), rearrange those pieces to make two balls each the same size as the original. One explicit construction uses five pieces. The result also holds in every dimension ≥ 3.
>
> [...]
>
> Summary
> - Banach–Tarski is a striking consequence of the Axiom of Choice and the non-amenability of 3D rotations: a ball can be split into finitely many non-measurable pieces and rigidly moved to yield two balls of the original size. It’s mathematically consistent and precise, but it applies to idealized sets, not to physical matter.

Between these parts the answer covers why volume is not conserved, the role of the Axiom of Choice, a proof sketch, why the result needs three dimensions, common misconceptions and a short reading list.

## Input Parameters
| Field | Type | Required | Default | Allowed values | Description |
| --- | --- | --- | --- | --- | --- |
| `prompt` | string | No | none |  | The prompt to send to the model. Do not use if using messages. |
| `system_prompt` | string | No | none |  | System prompt to set the assistant's behavior |
| `messages` | array of objects | No | `[]` |  | A list of messages, for example `[{"role": "user", "content": "Hello, how are you?"}]`. If provided, prompt and system_prompt are ignored. |
| `image_input` | array of URLs | No | `[]` |  | List of image URLs to send to the model |
| `verbosity` | string | No | `medium` | `low`, `medium`, `high` | Constrains the verbosity of the response. Lower values give more concise responses, higher values more comprehensive ones. |
| `max_completion_tokens` | integer | No | none |  | Maximum number of completion tokens to generate. For higher reasoning efforts you may need to increase your max_completion_tokens to avoid empty responses (where all the tokens are used on reasoning). |

## Pricing
- **Pay As You Go (PAYG) pricing**: you only pay for what you use.
- **Price**: $0.015 per API unit.
- **1 API unit = 100 weighted tokens**; an output token counts 1, an input token counts 0.125.
- That is **$18.75 per million input tokens** and **$150 per million output tokens**.
- Each request's weighted token total is rounded up to a whole unit.
- Output tokens include the tokens the model spends on reasoning, so the output count can be much larger than the visible answer.
- **Example cost**: the example request used 17 input tokens and 5,304 output tokens: 17 x 0.125 + 5,304 = 5,306.125 weighted tokens, rounded up to 54 units = **$0.81**.
- **Failed or cancelled predictions are not billed.**
- **GET requests to check prediction status** are free and don't count toward API unit usage.
- **No monthly subscription fee.**

## API Documentation

### Base URL
All API requests go to: `https://prod.api.market/api/v1/magicapi/ai-language-model-openai-gpt-5-pro`

### Authentication
Send your api.market API key in the `x-api-market-key` header:
```
x-api-market-key: YOUR_API_MARKET_KEY
```

### Asynchronous Processing
**IMPORTANT**: This API is asynchronous. The flow is:
1. Send a POST request to `/predictions`. A successful request returns **201 Created** with a prediction `id` and a `status` of `starting` or `processing`.
2. Poll GET `/predictions/{id}` with that ID.
3. Keep polling while `status` is `starting` or `processing`.
4. Stop when `status` is `succeeded` (read `output`), `failed` or `canceled` (read `error`).

The model reasons before it answers, so requests can take minutes: the example took 294 seconds from creation to completion. A polling interval of 2 seconds and an overall timeout of 1200 seconds (20 minutes) work well.

### Endpoints

#### Create a Prediction
```bash
curl -X POST "https://prod.api.market/api/v1/magicapi/ai-language-model-openai-gpt-5-pro/predictions" \
  -H "Content-Type: application/json" \
  -H "x-api-market-key: YOUR_API_MARKET_KEY" \
  -d '{
    "version": "openai/gpt-5-pro",
    "input": {
      "prompt": "Can you explain the Banach–Tarski paradox?",
      "messages": [],
      "verbosity": "medium",
      "image_input": []
    }
  }'
```

Response (201 Created):
```json
{
  "id": "8rgrqkj95xrmy0d1f4w81nj6g8",
  "version": "openai/gpt-5-pro",
  "status": "starting",
  "created_at": "2026-10-26T17:18:14.509Z"
}
```

#### Check Prediction Status
```bash
curl -X GET "https://prod.api.market/api/v1/magicapi/ai-language-model-openai-gpt-5-pro/predictions/{id}" \
  -H "x-api-market-key: YOUR_API_MARKET_KEY"
```

Response when the prediction has finished (200 OK, answer shortened):
```json
{
  "id": "8rgrqkj95xrmy0d1f4w81nj6g8",
  "status": "succeeded",
  "output": [
    "Here’s the idea, in layers.\n\nWhat the theorem says\n- In 3-dimensional space, you can cut a solid ball into finitely many disjoint pieces and, using only rotations and translations (no stretching), rearrange those pieces to make two balls each the same size as the original. One explicit construction uses five pieces. The result also holds in every dimension ≥ 3.\n..."
  ],
  "error": null,
  "created_at": "2026-10-26T17:18:14.509Z",
  "started_at": "2026-10-26T17:18:31.274Z",
  "completed_at": "2026-10-26T17:23:08.801Z"
}
```

`output` is a list of text chunks. Join them in order to get the full response. In the example the whole answer arrived as one chunk, shortened above.

### Status Codes
- **201 Created**: prediction created (POST `/predictions`). This is the normal success response.
- **200 OK**: prediction details returned (GET `/predictions/{id}`).
- **400 Bad Request**: invalid input, for example a `verbosity` value that is not in the list.
- **401 Unauthorized**: missing or invalid API key.
- **404 Not Found**: no prediction with that ID.
- **429 Too Many Requests**: rate limit exceeded.
- **500 Internal Server Error** or **503 Service Unavailable**: temporary server problem; retry later.

## Rate Limits
All plans allow 10 requests per second. Requests above that return `429 Too Many Requests`; back off before retrying. Polling a prediction with GET is free but still counts toward the per-second limit, so poll every 1 to 5 seconds rather than in a tight loop.

## Error Handling
- Check the HTTP status code of every request and read the response body for the error message.
- A prediction can also end with `status` set to `failed`; the `error` field then explains why. Failed predictions are not billed.
- Retry 429, 500 and 503 responses with exponential backoff. Fix the input before retrying a 400.

## Implementation Guide
### Example Code

#### JavaScript
```javascript
const apiKey = "YOUR_API_MARKET_KEY";
const baseUrl = "https://prod.api.market/api/v1/magicapi/ai-language-model-openai-gpt-5-pro";

async function createAndWaitForPrediction(input) {
  const response = await fetch(`${baseUrl}/predictions`, {
    method: "POST",
    headers: {
      "Content-Type": "application/json",
      "x-api-market-key": apiKey
    },
    body: JSON.stringify({ version: "openai/gpt-5-pro", input })
  });

  if (response.status !== 201) {
    throw new Error(`API error: ${response.status} ${await response.text()}`);
  }

  const prediction = await response.json();
  return pollForResults(prediction.id);
}

const delay = (ms) => new Promise((resolve) => setTimeout(resolve, ms));

async function pollForResults(predictionId) {
  const maxTimeout = 1200000;
  const pollInterval = 2000;
  const startTime = Date.now();

  while (Date.now() - startTime < maxTimeout) {
    const response = await fetch(`${baseUrl}/predictions/${predictionId}`, {
      headers: { "x-api-market-key": apiKey }
    });

    if (!response.ok) {
      throw new Error(`API error: ${response.status} ${await response.text()}`);
    }

    const prediction = await response.json();

    if (prediction.status === "succeeded") {
      return prediction.output.join("");
    }

    if (prediction.status === "failed" || prediction.status === "canceled") {
      throw new Error(`Prediction ${prediction.status}: ${prediction.error || "Unknown error"}`);
    }

    await delay(pollInterval);
  }

  throw new Error("Prediction timed out after 20 minutes");
}

createAndWaitForPrediction({
  "prompt": "Can you explain the Banach–Tarski paradox?",
  "messages": [],
  "verbosity": "medium",
  "image_input": []
}).then((text) => console.log(text));
```

#### Python
```python
import time
import requests

API_KEY = "YOUR_API_MARKET_KEY"
BASE_URL = "https://prod.api.market/api/v1/magicapi/ai-language-model-openai-gpt-5-pro"


def create_and_wait_for_prediction(input_data):
    headers = {"Content-Type": "application/json", "x-api-market-key": API_KEY}
    response = requests.post(
        f"{BASE_URL}/predictions",
        headers=headers,
        json={"version": "openai/gpt-5-pro", "input": input_data},
    )

    if response.status_code != 201:
        raise Exception(f"API error: {response.status_code} {response.text}")

    return poll_for_results(response.json()["id"])


def poll_for_results(prediction_id):
    max_timeout = 1200
    poll_interval = 2
    headers = {"x-api-market-key": API_KEY}
    start_time = time.time()

    while time.time() - start_time < max_timeout:
        response = requests.get(f"{BASE_URL}/predictions/{prediction_id}", headers=headers)

        if response.status_code != 200:
            raise Exception(f"API error: {response.status_code} {response.text}")

        prediction = response.json()

        if prediction["status"] == "succeeded":
            return "".join(prediction["output"])

        if prediction["status"] in ("failed", "canceled"):
            raise Exception(f"Prediction {prediction['status']}: {prediction.get('error') or 'Unknown error'}")

        time.sleep(poll_interval)

    raise Exception("Prediction timed out after 20 minutes")


print(create_and_wait_for_prediction({
    "prompt": "Can you explain the Banach–Tarski paradox?",
    "messages": [],
    "verbosity": "medium",
    "image_input": []
}))
```

## FAQ
- **Should I use `prompt` or `messages`?**
  - Use `prompt` (and `system_prompt`) for single questions. Use `messages` to send a whole conversation; when it is set, `prompt` and `system_prompt` are ignored.
- **Why did I get an empty answer?**
  - The model can use all of `max_completion_tokens` on reasoning before it writes the answer. Raise `max_completion_tokens` or leave it unset.
- **Why are output tokens so much higher than the length of the answer?**
  - Output tokens include the tokens the model spends on reasoning. In the example, a 3,733-character answer was billed as 5,304 output tokens.
- **How do I get the full answer text?**
  - Join the strings in `output` in order: `output.join("")` in JavaScript or `"".join(output)` in Python.
- **What did the example request cost?**
  - It used 17 input and 5,304 output tokens, which is 5,306.125 weighted tokens, rounded up to 54 units ($0.81).
- **Is the 201 response from POST an error?**
  - No. 201 Created means the prediction was accepted. Use the returned `id` to poll GET `/predictions/{id}` until `status` is `succeeded`, `failed` or `canceled`.

## Contact Us
If you have any questions or need assistance, you can reach out to us through the embedded chat on the bottom right corner of the page or send an email to hello@magicapi.com.


Send prompts, chats and images to OpenAI GPT-5 Pro, a reasoning model with verbosity control. $18.75/M input and $150/M output tokens, pay as you go.