Name: Qwen3-235B-A22B Instruct API: Open LLM by Alibaba
Brand: magicapi
SKU: api-magicapi-instruction-following-ai-api-qwen-qwen3-235b-a22b-instruct-2507
Availability: InStock
Author: API Market

Shashank Agarwal

# 🧠 Instruction-Following Language Model API (Alibaba Qwen Qwen3-235B-A22B-Instruct-2507)

## Overview
Qwen3-235B-A22B-Instruct-2507 is a large language model from Alibaba's Qwen team. Its developer describes it as an updated Qwen3 model for instruction following. It is a mixture-of-experts model with 235 billion total and 22 billion active parameters, and its weights are released under the Apache 2.0 license.

Send a text `prompt` and optionally set `max_tokens` and the sampling controls. The reply comes back as a list of text chunks; join them in order to get the full text. You pay per token: input and output tokens are counted separately.

## Features
- **Instruction following**: plans, explanations, tables and structured answers from one `prompt`.
- **Long replies**: `max_tokens` goes from 1 to 16384 (default 1024).
- **Low default temperature**: `temperature` defaults to 0.1 for focused, repeatable answers; raise it (up to 2) for more varied text.
- **Sampling controls**: `top_p` (0 to 1, default 1), `presence_penalty` and `frequency_penalty` (-2 to 2, default 0).
- **Open weights**: Apache 2.0 licensed model.

## Use Cases
1. **Plans and schedules**: training plans, study plans and project checklists formatted as Markdown tables.
2. **Question answering**: detailed explanations on technical and general topics.
3. **Writing**: drafts, rewrites and summaries that follow formatting instructions in the prompt.
4. **Structured output**: lists, tables and JSON-like text described in the prompt.

## Visual Examples
### Example Input
Prompt: "Give me a workout plan for this entire month to train for the Salkantay trek"

Sent with `max_tokens` `1024`, `temperature` `0.1`, `top_p` `1` and both penalties at `0`.

### Example Output
The output chunks joined in order (26 input tokens, 1024 output tokens). The reply stops mid-table because it reached `max_tokens` (1024):

```text
Absolutely! The Salkantay Trek is a challenging 5–7 day high-altitude trek in the Peruvian Andes, covering roughly 70–90 km (43–56 miles) with elevations reaching over 4,600 meters (15,090 ft). Proper training is essential to build endurance, strength, and mental resilience. Below is a **4-week progressive workout plan** designed to prepare you physically and mentally for the Salkantay Trek.

---

### 🎯 **Training Goals**
- Build cardiovascular endurance
- Strengthen legs and core
- Improve balance and stability
- Acclimate to elevation (as much as possible)
- Practice hiking with a loaded backpack

---

## 🗓️ **4-Week Training Plan Overview**

| Week | Focus |
|------|------|
| Week 1 | Foundation & Endurance |
| Week 2 | Strength & Elevation Simulation |
| Week 3 | Intensity & Long Hikes |
| Week 4 | Taper & Final Prep |

---

## 🏋️‍♂️ Weekly Structure (5–6 Days/Week)
- **Mon**: Cardio + Core
- **Tue**: Strength Training
- **Wed**: Active Recovery or Hike
- **Thu**: Cardio (Hill or Stairs)
- **Fri**: Strength + Balance
- **Sat**: Long Hike (Progressive)
- **Sun**: Rest or Light Walk

---

### ✅ **Week 1: Foundation & Endurance**
**Goal**: Build aerobic base and get used to consistent training.

| Day | Workout |
|-----|--------|
| Mon | 30-min brisk walk/jog + 15-min core (planks, crunches, leg raises) |
| Tue | Lower body strength: Squats 3x12, Lunges 3x10 each leg, Glute bridges 3x15 |
| Wed | 45-min hike or walk on incline (treadmill or hills) |
| Thu | 30-min stair climber or hill repeats (find steep hill, walk up fast, walk down, repeat 6x) |
| Fri | Strength: Step-ups 3x10, Calf raises 3x15, Side planks 3x30 sec |
| Sat | **Long Hike**: 2–3 hours with light backpack (5–10 lbs) |
| Sun | Rest or 30-min walk |

---

### ✅ **Week 2: Strength & Elevation Simulation**
**Goal**: Increase intensity and simulate trekking conditions.

| Day | Workout |
|-----|--------|
| Mon | 40-min jog or power walk + core circuit (3 rounds: 30-sec planks, 15 crunches, 15 leg raises) |
| Tue | Strength: Goblet squats 3x12, Bulgarian split squats 3x10, Deadlifts 3x10 |
| Wed | 60-min hike with 10–15 lb backpack |
| Thu | Stair climbing: 45 mins (or hill repeats x8) |
| Fri | Strength + balance: Step-ups with pack, single-leg stands, heel-toe walks |
| Sat | **Long Hike**: 3–4 hours with 15 lb backpack on hilly terrain |
| Sun | Rest or yoga/stretching |

---

### ✅ **Week 3: Intensity & Long Hikes**
**Goal**: Build stamina and simulate multi-day trekking.

| Day | Workout |
|-----|--------|
| Mon | 45-min cardio (brisk hike or elliptical) + core |
| Tue | Strength: Weighted lunges 3x12, wall sits 3x1 min, glute bridges with weight |
| Wed | 60-min hike with 15–20 lb backpack |
| Thu | Hill repeats: 8–10 rounds (walk up fast, walk down) |
| Fri | Balance & stability: Single-leg squats, Bosu ball or uneven surface walking |
| Sat | **Long Hike**: 5–6 hours with 20 lb backpack (simulate full day trekking) |
| Sun | Rest or light walk |

> 💡 *If possible, do this hike on varied terrain with elevation gain.*

---

### ✅ **Week 4: Taper & Final Prep**
**Goal**: Recover, reduce volume, and stay sharp.

| Day | Workout |
|-----|--------|
| Mon | 30-min easy walk + light core |
| Tue | Light strength: Bodyweight squats, lunges, step-ups (2x10) |
| Wed | 60-min easy hike with 10 lb backpack |
| Thu | 20-min stairs or incline walk (light effort)
```

## Input Parameters
| Field | Type | Required | Default | Allowed values | Description |
| --- | --- | --- | --- | --- | --- |
| `prompt` | string | No | `""` |  | The prompt to send to the model |
| `max_tokens` | integer | No | `1024` | `1` to `16384` | The maximum number of tokens the model should generate as output |
| `temperature` | number | No | `0.1` | `0` to `2` | Sampling temperature. Lower values make the output more deterministic |
| `presence_penalty` | number | No | `0` | `-2` to `2` | Presence penalty |
| `frequency_penalty` | number | No | `0` | `-2` to `2` | Frequency penalty |
| `top_p` | number | No | `1` | `0` to `1` | Top-p (nucleus) sampling |

## Pricing
- **Pay As You Go (PAYG) pricing**: you only pay for what you use.
- **Price**: $0.0014 per API unit.
- **1 API unit = 1,000 weighted tokens; an output token counts 1, an input token counts 0.2491.**
- That works out to **$0.34874 per million input tokens** and **$1.40 per million output tokens**.
- Each request is rounded up to a whole API unit, so every successful request costs at least $0.0014.
- **Example cost**: the example request used 26 input tokens and 1024 output tokens: 26 x 0.2491 + 1024 = 1030.4766 weighted tokens, rounded up to 2 API units = **$0.0028**.
- **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/instruction-following-ai-api-qwen-qwen3-235b-a22b-instruct-2507`

### 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`).

A polling interval of 2 seconds and an overall timeout of 900 seconds (15 minutes) work well. The example request, which generated 1024 tokens, took about 132 seconds.

### Endpoints

#### Create a Prediction
```bash
curl -X POST "https://prod.api.market/api/v1/magicapi/instruction-following-ai-api-qwen-qwen3-235b-a22b-instruct-2507/predictions" \
  -H "Content-Type: application/json" \
  -H "x-api-market-key: YOUR_API_MARKET_KEY" \
  -d '{
    "version": "qwen/qwen3-235b-a22b-instruct-2507",
    "input": {
      "top_p": 1,
      "prompt": "Give me a workout plan for this entire month to train for the Salkantay trek",
      "max_tokens": 1024,
      "temperature": 0.1,
      "presence_penalty": 0,
      "frequency_penalty": 0
    }
  }'
```

Response (201 Created):
```json
{
  "id": "hzqcvcydj9rmw0d1c6tbyyq0fr",
  "version": "qwen/qwen3-235b-a22b-instruct-2507",
  "status": "starting",
  "created_at": "2026-10-22T03:43:21.721Z"
}
```

#### Check Prediction Status
```bash
curl -X GET "https://prod.api.market/api/v1/magicapi/instruction-following-ai-api-qwen-qwen3-235b-a22b-instruct-2507/predictions/{id}" \
  -H "x-api-market-key: YOUR_API_MARKET_KEY"
```

Response when the prediction has finished (200 OK):
```json
{
  "id": "hzqcvcydj9rmw0d1c6tbyyq0fr",
  "status": "succeeded",
  "output": [
    "Absolutely! The Salkantay Trek is a challenging 5–7 day high-altitude trek in the Peruvian Andes, covering roughly 70–90 km (43–56 miles) with elevations reaching over 4,600 meters (15,090 ft). Proper training is essential to build endurance, strength, and mental resilience. Below is a **4-week progressive workout plan** designed to prepare you physically and mentally for the Salkantay Trek.\n\n---\n\n### 🎯 **Training Goals**\n- Build cardiovascular endurance\n- Strengthen legs and core\n- Improve balance and stability\n- Acclimate to elevation (as much as possible)\n- Practice hiking with a loaded backpack\n\n---\n\n## 🗓️ **4-Week Training Plan Overview**\n\n| Week | Focus |\n|------|------|\n| Week 1 | Foundation & Endurance |\n| Week 2 | Strength & Elevation Simulation |\n| Week 3 | Intensity & Long Hikes |\n| Week 4 | Taper & Final Prep |\n\n---\n\n## 🏋️‍♂️ Weekly Structure (5–6 Days/Week)\n- **Mon**: Cardio + Core\n- **Tue**: Strength Training\n- **Wed**: Active Recovery or Hike\n- **Thu**: Cardio (Hill or Stairs)\n- **Fri**: Strength + Balance\n- **Sat**: Long Hike (Progressive)\n- **Sun**: Rest or Light Walk\n\n---\n\n### ✅ **Week 1: Foundation & Endurance**\n**Goal**: Build aerobic base and get used to consistent training.\n\n| Day | Workout |\n|-----|--------|\n| Mon | 30-min brisk walk/jog + 15-min core (planks, crunches, leg raises) |\n| Tue | Lower body strength: Squats 3x12, Lunges 3x10 each leg, Glute bridges 3x15 |\n| Wed | 45-min hike or walk on incline (treadmill or hills) |\n| Thu | 30-min stair climber or hill repeats (find steep hill, walk up fast, walk down, repeat 6x) |\n| Fri | Strength: Step-ups 3x10, Calf raises 3x15, Side planks 3x30 sec |\n| Sat | **Long Hike**: 2–3 hours with light backpack (5–10 lbs) |\n| Sun | Rest or 30-min walk |\n\n---\n\n### ✅ **Week 2: Strength & Elevation Simulation**\n**Goal**: Increase intensity and simulate trekking conditions.\n\n| Day | Workout |\n|-----|--------|\n| Mon | 40-min jog or power walk + core circuit (3 rounds: 30-sec planks, 15 crunches, 15 leg raises) |\n| Tue | Strength: Goblet squats 3x12, Bulgarian split squats 3x10, Deadlifts 3x10 |\n| Wed | 60-min hike with 10–15 lb backpack |\n| Thu | Stair climbing: 45 mins (or hill repeats x8) |\n| Fri | Strength + balance: Step-ups with pack, single-leg stands, heel-toe walks |\n| Sat | **Long Hike**: 3–4 hours with 15 lb backpack on hilly terrain |\n| Sun | Rest or yoga/stretching |\n\n---\n\n### ✅ **Week 3: Intensity & Long Hikes**\n**Goal**: Build stamina and simulate multi-day trekking.\n\n| Day | Workout |\n|-----|--------|\n| Mon | 45-min cardio (brisk hike or elliptical) + core |\n| Tue | Strength: Weighted lunges 3x12, wall sits 3x1 min, glute bridges with weight |\n| Wed | 60-min hike with 15–20 lb backpack |\n| Thu | Hill repeats: 8–10 rounds (walk up fast, walk down) |\n| Fri | Balance & stability: Single-leg squats, Bosu ball or uneven surface walking |\n| Sat | **Long Hike**: 5–6 hours with 20 lb backpack (simulate full day trekking) |\n| Sun | Rest or light walk |\n\n> 💡 *If possible, do this hike on varied terrain with elevation gain.*\n\n---\n\n### ✅ **Week 4: Taper & Final Prep**\n**Goal**: Recover, reduce volume, and stay sharp.\n\n| Day | Workout |\n|-----|--------|\n| Mon | 30-min easy walk + light core |\n| Tue | Light strength: Bodyweight squats, lunges, step-ups (2x10) |\n| Wed | 60-min easy hike with 10 lb backpack |\n| Thu | 20-min stairs or incline walk (light effort)"
  ],
  "error": null,
  "created_at": "2026-10-22T03:43:21.721Z",
  "started_at": "2026-10-22T03:43:22.721Z",
  "completed_at": "2026-10-22T03:45:34.443Z"
}
```

`output` is a list of text chunks. Join them in order to get the full reply. The example above shows the reply joined into one element; a real response usually splits it into many short chunks.

### 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_tokens` above 16384 or a `temperature` above 2.
- **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 and honour any `Retry-After` header. 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/instruction-following-ai-api-qwen-qwen3-235b-a22b-instruct-2507";

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: "qwen/qwen3-235b-a22b-instruct-2507", 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 = 900000;
  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 15 minutes");
}

createAndWaitForPrediction({
  "top_p": 1,
  "prompt": "Give me a workout plan for this entire month to train for the Salkantay trek",
  "max_tokens": 1024,
  "temperature": 0.1,
  "presence_penalty": 0,
  "frequency_penalty": 0
}).then((output) => console.log(output.join("")));
```

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

API_KEY = "YOUR_API_MARKET_KEY"
BASE_URL = "https://prod.api.market/api/v1/magicapi/instruction-following-ai-api-qwen-qwen3-235b-a22b-instruct-2507"


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": "qwen/qwen3-235b-a22b-instruct-2507", "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 = 900
    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 15 minutes")


output = create_and_wait_for_prediction({
    "top_p": 1,
    "prompt": "Give me a workout plan for this entire month to train for the Salkantay trek",
    "max_tokens": 1024,
    "temperature": 0.1,
    "presence_penalty": 0,
    "frequency_penalty": 0
})
print("".join(output))
```

## FAQ
- **How do I get the full reply text?**
  - `output` is an array of text chunks. Join them in order with no separator: `output.join("")` in JavaScript or `"".join(output)` in Python.
- **Is there a system prompt or chat history field?**
  - No. The only text input is `prompt`. Put instructions and any earlier conversation at the start of `prompt`.
- **Why did the reply stop in the middle of a sentence?**
  - It reached `max_tokens`. The default is 1024; raise it (up to 16384) for long answers such as full monthly plans.
- **How much does a request cost?**
  - $0.34874 per million input tokens and $1.40 per million output tokens, rounded up to a whole API unit ($0.0014) per request. The example request cost $0.0028.
- **How long does a request take?**
  - It grows with the output length. The example request generated 1024 tokens in about 132 seconds, so allow a generous polling timeout for long replies.
- **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, reach out to us through the chat on the bottom right of the page. You can also email our support team at hello@magicapi.com.


Call Alibaba's Qwen3-235B-A22B-Instruct-2507 for instruction following, writing and planning from a text prompt. Pay as you go, no monthly fee.