---
title: "AI21"
url: "/docs/providers/ai21"
canonical_url: "https://docs.litellm.ai/docs/providers/ai21"
type: "docs"
last_updated: "2026-10-04"
summary: "LiteLLM supports the following AI21 models:"
related:
  - "/docs/providers/abliteration"
  - "/docs/providers/aiml"
---
# AI21

> Index of all LiteLLM docs: https://docs.litellm.ai/llms.txt


LiteLLM supports the following [AI21](https://www.ai21.com/studio/pricing) models:
* `jamba-1.5-mini`
* `jamba-1.5-large`
* `j2-light`
* `j2-mid`
* `j2-ultra`

:::tip

**We support ALL AI21 models, just set `model=ai21/<any-model-on-ai21>` as a prefix when sending litellm requests**. 
**See all litellm supported AI21 models [here](https://models.litellm.ai)**

:::

### API KEYS
```python
import os 
os.environ["AI21_API_KEY"] = "your-api-key"
```

## **LiteLLM Python SDK Usage**
### Sample Usage

```python
from litellm import completion 

# set env variable 
os.environ["AI21_API_KEY"] = "your-api-key"

messages = [{"role": "user", "content": "Write me a poem about the blue sky"}]

completion(model="ai21/jamba-1.5-mini", messages=messages)
```

## **LiteLLM Proxy Server Usage**

Here's how to call a ai21 model with the LiteLLM Proxy Server

1. Modify the config.yaml 

  ```yaml
  model_list:
    - model_name: my-model
      litellm_params:
        model: ai21/<your-model-name>  # add ai21/ prefix to route as ai21 provider
        api_key: api-key                 # api key to send your model
  ```

2. Start the proxy 

  ```bash
  $ litellm --config /path/to/config.yaml
  ```

3. Send Request to LiteLLM Proxy Server

**OpenAI Python v1.0.0+**

  ```python
  import openai
  client = openai.OpenAI(
      api_key="sk-<your-litellm-api-key>",             # pass litellm proxy key, if you're using virtual keys
      base_url="http://0.0.0.0:4000" # litellm-proxy-base url
  )

  response = client.chat.completions.create(
      model="my-model",
      messages = [
          {
              "role": "user",
              "content": "what llm are you"
          }
      ],
  )

  print(response)
  ```

**curl**

  ```shell
  curl --location 'http://0.0.0.0:4000/chat/completions' \
      --header "Authorization: Bearer $LITELLM_API_KEY" \
      --header 'Content-Type: application/json' \
      --data '{
      "model": "my-model",
      "messages": [
          {
          "role": "user",
          "content": "what llm are you"
          }
      ],
  }'
  ```

## Supported OpenAI Parameters

| [param](../completion/input) | type | AI21 equivalent |
|-------|-------------|------------------|
| `tools` | **Optional[list]** | `tools` |
| `response_format` | **Optional[dict]** | `response_format` |
| `max_tokens` | **Optional[int]** | `max_tokens` |
| `temperature` | **Optional[float]** | `temperature` |
| `top_p` | **Optional[float]** | `top_p` |
| `stop` | **Optional[Union[str, list]]** | `stop` |
| `n` | **Optional[int]** | `n` |
| `stream` | **Optional[bool]** | `stream` |
| `seed` | **Optional[int]** | `seed` |
| `tool_choice` | **Optional[str]** | `tool_choice` |
| `user` | **Optional[str]** | `user` |

## Supported AI21 Parameters

| param | type | [AI21 equivalent](https://docs.ai21.com/reference/jamba-15-api-ref#request-parameters) |
|-----------|------|-------------|
| `documents` | **Optional[List[Dict]]** | `documents` |

## Passing AI21 Specific Parameters -  `documents`

LiteLLM allows you to pass all AI21 specific parameters to the `litellm.completion` function. Here is an example of how to pass the `documents` parameter to the `litellm.completion` function.

**LiteLLM Python SDK**

```python
response = await litellm.acompletion(
    model="jamba-1.5-large",
    messages=[{"role": "user", "content": "what does the document say"}],
    documents = [
        {
            "content": "hello world",
            "metadata": {
                "source": "google",
                "author": "ishaan"
            }
        }
    ]
)

```

**LiteLLM Proxy Server**

```python
import openai
client = openai.OpenAI(
    api_key="sk-<your-litellm-api-key>",             # pass litellm proxy key, if you're using virtual keys
    base_url="http://0.0.0.0:4000" # litellm-proxy-base url
)

response = client.chat.completions.create(
    model="my-model",
    messages = [
        {
            "role": "user",
            "content": "what llm are you"
        }
    ],
    extra_body = {
        "documents": [
            {
                "content": "hello world",
                "metadata": {
                    "source": "google",
                    "author": "ishaan"
                }
            }
        ]
    }
)

print(response)

```

:::tip

**We support ALL AI21 models, just set `model=ai21/<any-model-on-ai21>` as a prefix when sending litellm requests**
**See all litellm supported AI21 models [here](https://models.litellm.ai)**
:::

## AI21 Models

| Model Name       | Function Call                              | Required OS Variables                |
|------------------|--------------------------------------------|--------------------------------------|
| jamba-1.5-mini         | `completion('jamba-1.5-mini', messages)`         | `os.environ['AI21_API_KEY']`         |
| jamba-1.5-large         | `completion('jamba-1.5-large', messages)`         | `os.environ['AI21_API_KEY']`         |
| j2-light         | `completion('j2-light', messages)`         | `os.environ['AI21_API_KEY']`         |
| j2-mid           | `completion('j2-mid', messages)`           | `os.environ['AI21_API_KEY']`         |
| j2-ultra         | `completion('j2-ultra', messages)`         | `os.environ['AI21_API_KEY']`         |

## Related pages

- [Abliteration](https://docs.litellm.ai/docs/providers/abliteration.md)
- [AI/ML API](https://docs.litellm.ai/docs/providers/aiml.md)