Name: Llama 2 13B API: Base Text Completion LLM by Meta
Brand: magicapi
SKU: api-magicapi-large-language-model-api-meta-llama-2-13b
Availability: InStock
Author: API Market

Shashank Agarwal

# 📝 Text Completion LLM API (Meta Llama 2 13B Base)

## Overview
Llama 2 13B is the base (not chat-tuned) version of Meta's 13-billion-parameter Llama 2 language model.

A base model continues text: send the start of a story, a document or a pattern in `prompt` and the model writes what comes next. It does not follow instructions the way a chat model does, so write prompts as text to be continued. `max_tokens` caps the length, and `temperature`, `top_p`, `top_k`, `stop_sequences` and the penalties control the output. The text comes back as a list of chunks that you join.

## Features
- **Text continuation**: the model continues whatever text you send in `prompt`.
- **Length controls**: `max_tokens` (default 512) and `min_tokens`; `stop_sequences` ends generation at the strings you list.
- **Sampling controls**: `temperature` 0 to 5 (default 0.7), `top_p` 0 to 1 (default 0.95) and `top_k` (default 0).
- **Repetition and length penalties**: `presence_penalty` and `length_penalty` (0 to 5, default 1).
- **Custom prompt template**: `prompt_template` wraps your prompt; it must contain `{prompt}`.
- **Repeatable runs**: set `seed` to reproduce a result.

## Use Cases
1. **Story and copy drafts**: give an opening line and let the model continue it.
2. **Pattern completion**: few-shot prompts that show a format, followed by a new case to complete.
3. **Text data generation**: varied sample text for testing and prototyping.
4. **Research and comparison**: compare a base model with its chat-tuned version on the same prompt.

## Visual Examples
### Example Input
Prompt: "Once upon a time a llama explored"

Settings: `max_new_tokens` 128 (the older name for `max_tokens`), `min_new_tokens` -1, `temperature` 0.75, `top_p` 0.9, `top_k` 50.

### Example Output
The chunks of this run joined into one text (the model continues the prompt, so the output starts mid-sentence):

```text
the Andes. He was a very happy llama. He was curious about the world. He would see a tree or a rock and wonder what it was like inside. He would see a mountain and wonder what it was like on top. He would see a river and wonder what it was like to swim in it. He would see a bird and wonder what it was like to fly. He was so curious that he set off to explore the world.
The llama explored the Andes. He climbed mountains. He swam in rivers. He flew like a bird. He even explored the sky.
```

## Input Parameters
| Field | Type | Required | Default | Allowed values | Description |
| --- | --- | --- | --- | --- | --- |
| `prompt` | string | Yes |  |  | Prompt to send to the model |
| `max_tokens` | integer | No | `512` | 1 or more | Maximum number of tokens to generate. A word is generally 2 to 3 tokens. |
| `min_tokens` | integer | No | none | -1 or more | Minimum number of tokens to generate. Set to -1 to disable. A word is generally 2 to 3 tokens. |
| `temperature` | number | No | `0.7` | 0 to 5 | Randomness of the output: 0 is deterministic, above 1 is very random; 0.75 is a good starting value |
| `top_p` | number | No | `0.95` | 0 to 1 | Samples from the top p share of most likely tokens; lower it to ignore less likely tokens |
| `top_k` | integer | No | `0` | -1 or more | Samples from the top k most likely tokens; lower it to ignore less likely tokens |
| `stop_sequences` | string | No | none |  | Comma-separated list of sequences to stop generation at, for example '<end>,<stop>' |
| `length_penalty` | number | No | `1` | 0 to 5 | Controls output length: below 1 favors shorter outputs, above 1 favors longer outputs |
| `presence_penalty` | number | No | `0` |  | Penalizes repeated tokens regardless of how often they appear; higher values make repetition less likely |
| `seed` | integer | No | none |  | Random seed. Leave blank to randomize. |
| `prompt_template` | string | No | `{prompt}` |  | Template for formatting the prompt. Any string that contains the substring {prompt}. |
| `log_performance_metrics` | boolean | No | `false` |  | Log performance metrics for the request |
| `max_new_tokens` | integer | No | none | 1 or more | Older name for max_tokens, kept for backwards compatibility. Use max_tokens; do not send both. |
| `min_new_tokens` | integer | No | none | -1 or more | Older name for min_tokens, kept for backwards compatibility. Use min_tokens; do not send both. |

## Pricing
- **Pay As You Go (PAYG) pricing**: you only pay for what you use.
- **Price**: $0.001 per API unit.
- **1 API unit = 1,000 weighted tokens**; an output token counts 1, an input token counts 0.2.
- That is **$0.2 per million input tokens** and **$1 per million output tokens**.
- Each request is rounded up to a whole API unit, so the smallest charge per request is $0.001.
- **Example cost**: the request above used 10 input tokens and 128 output tokens. 10 x 0.2 + 128 = 130 weighted tokens, rounded up to 1 API unit = **$0.001**.
- **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/large-language-model-api-meta-llama-2-13b`

### 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.

### Endpoints

#### Create a Prediction
```bash
curl -X POST "https://prod.api.market/api/v1/magicapi/large-language-model-api-meta-llama-2-13b/predictions" \
  -H "Content-Type: application/json" \
  -H "x-api-market-key: YOUR_API_MARKET_KEY" \
  -d '{
    "version": "meta/llama-2-13b",
    "input": {
      "top_k": 50,
      "top_p": 0.9,
      "prompt": "Once upon a time a llama explored",
      "temperature": 0.75,
      "max_new_tokens": 128,
      "min_new_tokens": -1
    }
  }'
```

Response (201 Created):
```json
{
  "id": "0y8hp24mjhrna0d1d20b08087c",
  "version": "meta/llama-2-13b",
  "status": "starting",
  "created_at": "2026-10-23T11:23:39.749Z"
}
```

#### Check Prediction Status
```bash
curl -X GET "https://prod.api.market/api/v1/magicapi/large-language-model-api-meta-llama-2-13b/predictions/{id}" \
  -H "x-api-market-key: YOUR_API_MARKET_KEY"
```

Response when the prediction has finished (200 OK):
```json
{
  "id": "0y8hp24mjhrna0d1d20b08087c",
  "status": "succeeded",
  "output": [
    "the Andes. He was a very happy llama. He was curious about the world. He would see a tree or a rock and wonder what it was like inside. He would see a mountain and wonder what it was like on top. He would see a river and wonder what it was like to swim in it. He would see a bird and wonder what it was like to fly. He was so curious that he set off to explore the world.\nThe llama explored the Andes. He climbed mountains. He swam in rivers. He flew like a bird. He even explored the sky."
  ],
  "error": null,
  "created_at": "2026-10-23T11:23:39.749Z",
  "started_at": "2026-10-23T11:26:07.154Z",
  "completed_at": "2026-10-23T11:26:09.052Z"
}
```

`output` is an array of text chunks. Join them in order to get the full text. The example above shows the chunks from this run already joined into one string.

### 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` or a `temperature` above 5.
- **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/large-language-model-api-meta-llama-2-13b";

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: "meta/llama-2-13b", 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;
    }

    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({
    "top_k": 50,
    "top_p": 0.9,
    "prompt": "Once upon a time a llama explored",
    "temperature": 0.75,
    "max_new_tokens": 128,
    "min_new_tokens": -1
  }).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/large-language-model-api-meta-llama-2-13b"


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": "meta/llama-2-13b", "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 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("".join(create_and_wait_for_prediction({
    "top_k": 50,
    "top_p": 0.9,
    "prompt": "Once upon a time a llama explored",
    "temperature": 0.75,
    "max_new_tokens": 128,
    "min_new_tokens": -1
})))
```

## FAQ
- **Is this a chat model?**
  - No. It is the base model and continues text. For instruction following and conversation, use the Llama 2 13B Chat listing.
- **How do I stop the output at a certain point?**
  - List the strings to stop at in `stop_sequences`, separated by commas, or cap the length with `max_tokens`.
- **Should I use `max_tokens` or `max_new_tokens`?**
  - Use `max_tokens`. `max_new_tokens` is the older name and still works, but do not send both.
- **How much does a typical request cost?**
  - Each request is billed in units of 1,000 weighted tokens at $0.001 per unit. The example (10 input and 128 output tokens) cost $0.001.
- **How do I read the output?**
  - `output` is a list of text chunks. Join them in order, for example `output.join("")` in JavaScript or `"".join(output)` in Python.

## 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.


Generate text continuations with Meta's Llama 2 13B base model, with full control over sampling, stop sequences and prompt template.