Name: Pruna AI Qwen3-32B Text Generation LLM API
Brand: magicapi
SKU: api-magicapi-fast-ai-inference-prunaai-qwen3-32b
Availability: InStock
Author: API Market

Shashank Agarwal

# 💬 Text Generation LLM API (Pruna AI Qwen3-32B)

## Overview
Qwen3-32B is a 32-billion-parameter large language model from Alibaba's Qwen team. This listing runs a version optimized by Pruna AI with its open-source compression tooling; its developer describes it as about 2x faster than the original model.

Send a text `prompt` and the model returns its answer as one text string. Set `max_new_tokens` to cap the answer length and `enable_thinking` to choose whether the model writes out its intermediate reasoning before the answer.

## Features
- **Text in, text out**: one `prompt` string per request; the answer comes back as a single string in `output`.
- **Thinking mode**: `enable_thinking` (default `true`) makes the model output its intermediate reasoning. Set it to `false` for a direct answer.
- **Answer length control**: `max_new_tokens` (default `512`) sets the maximum number of new tokens the model generates.
- **Markdown answers**: the model formats answers with headings and lists when the prompt asks for structured content, as in the example below.

## Use Cases
1. **Drafting and writing**: checklists, outlines, short guides and summaries from a one-line instruction.
2. **Question answering**: direct answers to general knowledge questions with `enable_thinking` set to `false`.
3. **Step-by-step reasoning**: leave `enable_thinking` on when you want to see how the model works through a problem.
4. **Content for apps**: generate text that your application shows or post-processes, such as tips, descriptions or FAQ drafts.

## Visual Examples
### Example Input
```json
{
  "prompt": "Write a short checklist for preparing a small vegetable garden in spring.",
  "max_new_tokens": 2048,
  "enable_thinking": false
}
```

### Example Output
The `output` string, rendered as Markdown:

```text
Sure! Here's a short checklist to help prepare a small vegetable garden in spring:

### Spring Vegetable Garden Prep Checklist

#### 1. **Plan Your Garden**
- ✅ Decide on vegetables to grow (based on climate and space).
- ✅ Sketch layout or designate garden area.
- ✅ Check planting dates for your region.

#### 2. **Prepare the Soil**
- ✅ Clear weeds and debris from the garden area.
- ✅ Loosen soil with a spade or garden fork.
- ✅ Add compost or organic matter to enrich soil.
- ✅ Test soil pH (optional) and amend if necessary.

#### 3. **Gather Supplies**
- ✅ Seeds or seedlings.
- ✅ Garden tools (spade, trowel, rake).
- ✅ Mulch or straw (optional).
- ✅ Fertilizer (organic or balanced type).
- ✅ Watering can or hose with nozzle.

#### 4. **Planting**
- ✅ Start cool-season crops early (e.g., lettuce, peas, spinach).
- ✅ Sow seeds or transplant seedlings according to spacing and depth guidelines.
- ✅ Mark rows or beds with labels for easy identification.

#### 5. **Maintenance Begins**
- ✅ Water regularly, especially after planting.
- ✅ Monitor for pests and diseases.
- ✅ Mulch to retain moisture and suppress weeds.
- ✅ Keep a garden journal or planting schedule.

#### Bonus Tips:
- ✅ Use row covers or netting to protect young plants from pests.
- ✅ Rotate crops each year if possible to avoid soil depletion.

Let me know if you'd like a printable version or one tailored to a specific climate! 🌱
```

## Input Parameters
| Field | Type | Required | Default | Allowed values | Description |
| --- | --- | --- | --- | --- | --- |
| `prompt` | string | No | `How many helicopters can a human eat in one sitting?` |  | Prompt for the model |
| `max_new_tokens` | integer | No | `512` |  | Maximum number of new tokens to generate |
| `enable_thinking` | boolean | No | `true` |  | Enable thinking mode (outputs intermediate reasoning) |

`prompt` has a default in the schema, so always send your own prompt.

## Pricing
- **Pay As You Go (PAYG) pricing**: you only pay for what you use.
- **Price**: $0.064 per API unit.
- **1 API unit = 1 request with up to 41 seconds of processing.** A request that processes for longer bills one more unit for each started 41-second block (for example, 50 seconds of processing is 2 units, $0.128).
- Processing time grows with the length of the answer, so a higher `max_new_tokens` and `enable_thinking` set to `true` can raise the cost of a request.
- **Example cost**: the example request above processed for 180.2 seconds, which is 5 started 41-second blocks: 5 units, $0.32.
- **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/fast-ai-inference-prunaai-qwen3-32b`

### 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`).

Long answers can take several minutes. A polling interval of 2 seconds and an overall timeout of 1800 seconds (30 minutes) work well.

### Endpoints

#### Create a Prediction
```bash
curl -X POST "https://prod.api.market/api/v1/magicapi/fast-ai-inference-prunaai-qwen3-32b/predictions" \
  -H "Content-Type: application/json" \
  -H "x-api-market-key: YOUR_API_MARKET_KEY" \
  -d '{
    "version": "5ea670f086464b322044e36d46f65506db15334e7d777efc0e886162684d4831",
    "input": {
      "prompt": "Write a short checklist for preparing a small vegetable garden in spring.",
      "max_new_tokens": 2048,
      "enable_thinking": false
    }
  }'
```

Response (201 Created):
```json
{
  "id": "4hpmc0e205rne0d1mvns9sw6tg",
  "version": "5ea670f086464b322044e36d46f65506db15334e7d777efc0e886162684d4831",
  "status": "starting",
  "created_at": "2026-11-04T14:17:10.913Z"
}
```

#### Check Prediction Status
```bash
curl -X GET "https://prod.api.market/api/v1/magicapi/fast-ai-inference-prunaai-qwen3-32b/predictions/{id}" \
  -H "x-api-market-key: YOUR_API_MARKET_KEY"
```

Response when the prediction has finished (200 OK, `output` shortened here):
```json
{
  "id": "4hpmc0e205rne0d1mvns9sw6tg",
  "status": "succeeded",
  "output": "Sure! Here's a short checklist to help prepare a small vegetable garden in spring:\n\n### Spring Vegetable Garden Prep Checklist\n\n#### 1. **Plan Your Garden**\n- ✅ Decide on vegetables to grow (based on climate and space).\n- ✅ Sketch layout or designate garden area.\n- ✅ Check planting dates for your region.\n\n...\n\nLet me know if you'd like a printable version or one tailored to a specific climate! 🌱",
  "error": null,
  "created_at": "2026-11-04T14:17:10.913Z",
  "started_at": "2026-11-04T14:23:50.965Z",
  "completed_at": "2026-11-04T14:26:51.178Z"
}
```

`output` is a single string with the model's answer. With `enable_thinking` set to `true` it also contains the model's intermediate reasoning.

### 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 `max_new_tokens` that is not an integer.
- **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
Requests are rate limited to keep usage fair. If you receive a 429 response, wait and retry with exponential backoff.

## 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/fast-ai-inference-prunaai-qwen3-32b";

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: "5ea670f086464b322044e36d46f65506db15334e7d777efc0e886162684d4831", 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 = 1800000;
  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;
    }

    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 30 minutes");
}

createAndWaitForPrediction({
  "prompt": "Write a short checklist for preparing a small vegetable garden in spring.",
  "max_new_tokens": 2048,
  "enable_thinking": false
}).then((output) => console.log(output));
```

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

API_KEY = "YOUR_API_MARKET_KEY"
BASE_URL = "https://prod.api.market/api/v1/magicapi/fast-ai-inference-prunaai-qwen3-32b"


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": "5ea670f086464b322044e36d46f65506db15334e7d777efc0e886162684d4831", "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 = 1800
    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 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 30 minutes")


print(create_and_wait_for_prediction({
    "prompt": "Write a short checklist for preparing a small vegetable garden in spring.",
    "max_new_tokens": 2048,
    "enable_thinking": False
}))
```

## FAQ
- **How do I turn off the reasoning text?**
  - Set `enable_thinking` to `false`. With the default `true`, the model writes out its intermediate reasoning in `output` before the answer.
- **How long can an answer be?**
  - Up to `max_new_tokens` new tokens (default 512). Raise it for longer answers; the example used 2048.
- **What does one request cost?**
  - $0.064 for each started 41 seconds of processing, with a minimum of one unit. The example request processed for 180.2 seconds and cost $0.32 (5 units).
- **What format is the output?**
  - A single text string. The model often uses Markdown headings and lists, so render it as Markdown or show it as plain text.
- **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`.
- **What if my prediction fails?**
  - Failed predictions are not billed. Read the `error` field, adjust the input and try again.

## Contact Us
If you have any questions, reach out to us through the chat on the bottom right of the page.


Run Alibaba's Qwen3-32B large language model, optimized by Pruna AI, for text generation with optional thinking mode. $0.064 per 41 seconds of processing, pay as you go.