Name: Claude Opus 5.5 API: Agentic Coding and Knowledge Work
Brand: magicapi
SKU: api-magicapi-agentic-coding-knowledge-api-anthropic-claude-opus-5-5
Availability: InStock
Author: API Market

Shashank Agarwal

# 🧠 LLM Chat & Coding API (Anthropic Claude Opus 5.5)

## Overview
Claude Opus 5.5 is a large language model from Anthropic. Its developer describes it as an Opus model for long-running agentic coding and knowledge work, at a lower price than Opus 5.

Send a text `prompt`, with an optional `system_prompt` and an optional `image`, and the model returns its answer as a list of text chunks. The model always thinks before it answers; `effort` sets how much. Join the chunks to get the full answer.

## Features
- **Text in, text out**: send a `prompt` and read the answer from `output`.
- **System prompt**: set instructions, a persona or an output format with `system_prompt`.
- **Image input**: attach one `image` URL. The image is scaled down to at most `max_image_resolution` megapixels (default `0.5`, range 0.001 to 2) before it is sent to the model. An image counts as (width px x height px) / 750 input tokens.
- **Effort levels**: `low` (default, fastest and cheapest), `medium`, `high`, `xhigh`, `max`. Higher effort spends more output tokens on thinking and is slower.
- **Output limit**: `max_tokens` caps output tokens, thinking included. Minimum 1024, maximum 128000, default 8192.

## Use Cases
1. **Coding**: write, explain and review functions, scripts and tests.
2. **Agentic and multi-step tasks**: plan and work through long tasks at a higher `effort`.
3. **Knowledge work**: summaries, analysis, drafts and structured answers from long instructions.
4. **Questions about an image**: describe a screenshot, a chart or a photo sent as `image`.

## Visual Examples
### Example Input
```json
{
  "effort": "medium",
  "prompt": "Write a Python function that returns the longest palindromic substring of a string. Explain the approach briefly, then give the code with type hints.",
  "max_tokens": 8192
}
```

### Example Output
The chunks joined together (shortened):

```markdown
## Approach: Expand Around Center

Every palindrome mirrors around its center. A center is either:
- **a single character** (odd-length palindromes like `"aba"`), or
- **the gap between two characters** (even-length palindromes like `"abba"`).

A string of length *n* has `2n - 1` possible centers. For each one, expand outward while the characters on both sides match, and keep track of the longest palindrome found.

- **Time:** O(n²). There are O(n) centers, and each expansion costs up to O(n).
- **Space:** O(1) extra, apart from the returned substring.

## Code

def longest_palindrome(s: str) -> str:
    """Return the longest palindromic substring of s. ..."""
    if len(s) < 2:
        return s
    ...
```

The full answer also contains the complete function, a small test block with its printed results, and a note on Manacher's algorithm for O(n) time.

## Input Parameters
| Field | Type | Required | Default | Allowed values | Description |
| --- | --- | --- | --- | --- | --- |
| `prompt` | string | Yes |  |  | Input prompt |
| `image` | string (URL) | No | none |  | Optional input image. Images are priced as (width px x height px) / 750 input tokens |
| `system_prompt` | string | No | `""` |  | System prompt |
| `max_tokens` | integer | No | `8192` | 1024 to 128000 | Maximum number of output tokens, including thinking tokens. The model always thinks before responding, so leave headroom at higher effort levels (Anthropic recommends starting at 64000 for `xhigh` and `max`) |
| `effort` | string | No | `low` | `low`, `medium`, `high`, `xhigh`, `max` | How much thinking the model does before responding. Higher effort improves quality on hard coding and agentic tasks, but is slower and uses more output tokens. `low` is the fastest and cheapest |
| `max_image_resolution` | number | No | `0.5` | 0.001 to 2 | Maximum image resolution in megapixels. The image is scaled down before it is sent to the model, to save time and money |

## Pricing
- **Pay As You Go (PAYG) pricing**: you only pay for what you use.
- **Price**: $0.0025 per API unit.
- **1 API unit = 100 weighted tokens; an output token counts 1, an input token counts 0.2.**
- That works out to **$5 per million input tokens** and **$25 per million output tokens**. Thinking tokens are output tokens. Each request rounds up to a whole unit.
- An image adds (width px x height px) / 750 input tokens after it is scaled down to `max_image_resolution`.
- **Example cost**: the example request measured 49 input tokens and 1050 output tokens (thinking included). That is 1050 + 49 x 0.2 = 1059.8 weighted tokens, rounded up to 11 units, so $0.0275. Output length, and so the price, varies from run to run and grows with `effort`.
- **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/agentic-coding-knowledge-api-anthropic-claude-opus-5-5`

### 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 1 second and an overall timeout of 600 seconds (10 minutes) work well. Long answers at `xhigh` or `max` effort take longer.

### Endpoints

#### Create a Prediction
```bash
curl -X POST "https://prod.api.market/api/v1/magicapi/agentic-coding-knowledge-api-anthropic-claude-opus-5-5/predictions" \
  -H "Content-Type: application/json" \
  -H "x-api-market-key: YOUR_API_MARKET_KEY" \
  -d '{
    "version": "anthropic/claude-opus-5.5",
    "input": {
      "effort": "medium",
      "prompt": "Write a Python function that returns the longest palindromic substring of a string. Explain the approach briefly, then give the code with type hints.",
      "max_tokens": 8192
    }
  }'
```

Response (201 Created):
```json
{
  "id": "4wxt1h9z7hrmt0d1d92s3fva9r",
  "version": "anthropic/claude-opus-5.5",
  "status": "starting",
  "created_at": "2026-10-23T19:38:05.667Z"
}
```

#### Check Prediction Status
```bash
curl -X GET "https://prod.api.market/api/v1/magicapi/agentic-coding-knowledge-api-anthropic-claude-opus-5-5/predictions/{id}" \
  -H "x-api-market-key: YOUR_API_MARKET_KEY"
```

Response when the prediction has finished (200 OK, `output` shortened to its first chunks):
```json
{
  "id": "4wxt1h9z7hrmt0d1d92s3fva9r",
  "status": "succeeded",
  "output": [
    "## Approach:",
    " Expand Around Center",
    "\n\nEvery palindrome m",
    "irrors around its center. A",
    " center is",
    " either:"
  ],
  "error": null,
  "created_at": "2026-10-23T19:38:05.667Z",
  "started_at": "2026-10-23T19:38:06.667Z",
  "completed_at": "2026-10-23T19:38:16.667Z"
}
```

`output` is an array of text chunks. Join them in order, with no separator, to get the full answer (Markdown text).

### 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 missing `prompt`, a `max_tokens` below 1024 or an `effort` 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 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/agentic-coding-knowledge-api-anthropic-claude-opus-5-5";

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: "anthropic/claude-opus-5.5", 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 = 600000;
  const pollInterval = 1000;
  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 10 minutes");
}

createAndWaitForPrediction({
  "effort": "medium",
  "prompt": "Write a Python function that returns the longest palindromic substring of a string. Explain the approach briefly, then give the code with type hints.",
  "max_tokens": 8192
}).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/agentic-coding-knowledge-api-anthropic-claude-opus-5-5"


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": "anthropic/claude-opus-5.5", "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 = 600
    poll_interval = 1
    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 10 minutes")


print(create_and_wait_for_prediction({
    "effort": "medium",
    "prompt": "Write a Python function that returns the longest palindromic substring of a string. Explain the approach briefly, then give the code with type hints.",
    "max_tokens": 8192
}))
```

## FAQ
- **How do I get the full answer from `output`?**
  - `output` is a list of text chunks. Join them in order with no separator.
- **Which `effort` should I use?**
  - `low` (the default) is the fastest and cheapest. Use `medium` or `high` for harder coding and reasoning tasks, and `xhigh` or `max` for the hardest ones. Higher effort uses more output tokens, so it costs more.
- **What should I set `max_tokens` to?**
  - At least 1024 (the minimum) and at most 128000. Thinking tokens count toward the limit, so Anthropic recommends starting at 64000 for `xhigh` and `max`. The default is 8192.
- **How much does an image cost?**
  - An image counts as (width px x height px) / 750 input tokens after it is scaled down to `max_image_resolution` megapixels. At the default of 0.5 megapixels that is about 667 input tokens, roughly $0.0033.
- **Can I send more than one image or a chat history?**
  - No. The input takes one `prompt`, one optional `system_prompt` and one optional `image`. To continue a conversation, include the earlier turns in `prompt`.
- **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. You can also email our support team at hello@magicapi.com.


Run Anthropic Claude Opus 5.5 for coding, agentic tasks and knowledge work, with a system prompt, image input and five effort levels.