Name: Llama 2 7B API: Base Text Completion LLM by Meta
Brand: magicapi
SKU: api-magicapi-text-completion-meta-llama-2-7b
Availability: InStock
Author: API Market

Shashank Agarwal

# 💬 Text Completion API (Meta Llama 2 7B)

## Overview
Llama 2 7B is the base version of Llama 2 with 7 billion parameters, a language model from Meta.

It is a base model: it continues the text you send in `prompt` rather than following instructions or holding a conversation. Write the start of the text you want, or a few examples of the pattern, and the model carries on from there. The output comes back as a list of text chunks that you join in order.

## Features
- **Text continuation**: send any text in `prompt` and the model writes what comes next.
- **Length control**: `max_tokens` (default 512) caps the output and `min_tokens` sets a minimum. `length_penalty` (0 to 5, default 1) nudges outputs shorter or longer.
- **Sampling controls**: `temperature` (0 to 5, default 0.7), `top_p` (0 to 1, default 0.95) and `top_k` (default 0).
- **Stop sequences**: `stop_sequences` takes a comma-separated list of strings that end the output.
- **Prompt template**: `prompt_template` (default `{prompt}`) wraps your prompt, for example to add a fixed prefix.
- **Repetition control**: `presence_penalty` (default 0) makes repeated tokens less likely.
- **Reproducible runs**: set `seed` to repeat a result.
- **Chunked output**: the response is a list of text chunks that you join in order.

## Use Cases
1. **Short creative text**: poems, slogans and short descriptions started from a one-line prompt, as in the example.
2. **Few-shot generation**: show a few examples of a pattern in the prompt and let the model write the next one.
3. **Low-cost drafting**: generate first drafts or variations in bulk at a low price per token.
4. **Model comparison**: compare raw base-model completions with instruction-tuned models on the same prompts.

## Examples
### Example Input
A request for a short haiku about autumn leaves, with `max_new_tokens` 500, `temperature` 0.95 and `top_k` 250. A base model does not stop on its own after the poem, so it keeps writing until the token limit; set `max_tokens` lower or use `stop_sequences` to end earlier.

```json
{
  "top_k": 250,
  "top_p": 0.95,
  "prompt": "A short haiku about autumn leaves:\n",
  "temperature": 0.95,
  "max_new_tokens": 500,
  "min_new_tokens": -1
}
```

### Example Output
The joined output text, shortened here (the full response was 500 tokens):

~~~text
As they fall to the ground
Under the autumn sky,
The colors change, and die ...
~~~

## 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 it to -1 to disable |
| `temperature` | number | No | `0.7` | `0` to `5` | Randomness of the output: 0 is deterministic and values above 1 are very random. 0.75 is a good starting value |
| `top_p` | number | No | `0.95` | `0` to `1` | Samples only from the most likely tokens that together make up the top p probability; lower it to ignore less likely tokens |
| `top_k` | integer | No | `0` | `-1` or more | Samples only from the k most likely tokens; lower it to ignore less likely tokens |
| `stop_sequences` | string | No | none |  | Comma-separated list of sequences that stop generation, for example `<end>,<stop>` stops at the first `<end>` or `<stop>` |
| `length_penalty` | number | No | `1` | `0` to `5` | Controls output length: below 1 the model tends to write shorter outputs, above 1 longer ones |
| `presence_penalty` | number | No | `0` |  | Penalizes tokens that already appeared, however often; higher values make repetition less likely |
| `seed` | integer | No | none |  | Random seed. Leave it out to use a random seed |
| `prompt_template` | string | No | `{prompt}` |  | Template for formatting the prompt. Any string that contains `{prompt}` |
| `log_performance_metrics` | boolean | No | `false` |  | Debug option that logs performance metrics; it does not change the output |
| `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**: $0.0032 per API unit.
- **1 API unit = 10,000 weighted tokens**; an output token counts 1, an input token counts 0.2.
- That is **$0.064 per million input tokens** and **$0.32 per million output tokens**.
- Each request is rounded up to a whole API unit, so the smallest charge for a request is $0.0032.
- **Example cost**: the example request used 11 input tokens and 500 output tokens: 11 x 0.2 + 500 x 1 = 502.2 weighted tokens, rounded up to 1 API unit = **$0.0032**.
- Output tokens make up most of the cost, so `max_tokens` is the main way to cap the price of a request.
- **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/text-completion-meta-llama-2-7b`

### 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/text-completion-meta-llama-2-7b/predictions" \
  -H "Content-Type: application/json" \
  -H "x-api-market-key: YOUR_API_MARKET_KEY" \
  -d '{
    "version": "meta/llama-2-7b",
    "input": {
      "top_k": 250,
      "top_p": 0.95,
      "prompt": "A short haiku about autumn leaves:\n",
      "temperature": 0.95,
      "max_new_tokens": 500,
      "min_new_tokens": -1
    }
  }'
```

Response (201 Created):
```json
{
  "id": "jv2r8306vxrna0d1d20v1xnpmg",
  "version": "meta/llama-2-7b",
  "status": "starting",
  "created_at": "2026-10-23T11:24:09.076Z"
}
```

#### Check Prediction Status
```bash
curl -X GET "https://prod.api.market/api/v1/magicapi/text-completion-meta-llama-2-7b/predictions/{id}" \
  -H "x-api-market-key: YOUR_API_MARKET_KEY"
```

Response when the prediction has finished (200 OK):
```json
{
  "id": "jv2r8306vxrna0d1d20v1xnpmg",
  "status": "succeeded",
  "output": ["As", " they", " fall", " to", " the", " ground", "\n", "Under", " the", " aut", "umn", " sky", ",", "\n", "The", " colors", " change", ",", " and", " die"],
  "error": null,
  "created_at": "2026-10-23T11:24:09.076Z",
  "started_at": "2026-10-23T11:26:09.916Z",
  "completed_at": "2026-10-23T11:26:13.255Z"
}
```

`output` is a list of text chunks in the order the model produced them (shortened above). Join them with no separator to get the full 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 `temperature` above 5 or both `max_tokens` and `max_new_tokens` in one request.
- **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/text-completion-meta-llama-2-7b";

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-2-7b", 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": 250,
  "top_p": 0.95,
  "prompt": "A short haiku about autumn leaves:\n",
  "temperature": 0.95,
  "max_new_tokens": 500,
  "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/text-completion-meta-llama-2-7b"


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-2-7b", "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")


output = create_and_wait_for_prediction({
    "top_k": 250,
    "top_p": 0.95,
    "prompt": "A short haiku about autumn leaves:\n",
    "temperature": 0.95,
    "max_new_tokens": 500,
    "min_new_tokens": -1
})
print("".join(output))
```

## FAQ
- **How do I get the full text as one string?**
  - `output` is a list of text chunks. Join them in order with no separator: `output.join("")` in JavaScript or `"".join(output)` in Python.
- **Why does the model not answer my question directly?**
  - Llama 2 7B is a base model that continues text. Phrase the prompt as the beginning of the text you want (for example `Question: ... Answer:`), or use a chat or instruction-tuned model for direct answers.
- **How do I stop the model from running on?**
  - Lower `max_tokens`, or set `stop_sequences` to a string that marks the end of what you want, such as a blank line or a label your prompt uses.
- **Should I use `max_tokens` or `max_new_tokens`?**
  - Use `max_tokens`. `max_new_tokens` and `min_new_tokens` are older names kept for backwards compatibility; do not send both forms in one request.
- **How much does a request cost?**
  - $0.064 per million input tokens and $0.32 per million output tokens, billed in units of $0.0032 and rounded up per request. The example cost $0.0032.
- **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.


Generate text continuations and few-shot completions with Meta's 7-billion-parameter Llama 2 base model. Pay as you go, no monthly fee.