> For clean Markdown of any page, append `.md` to the page URL.
> For a complete documentation index, see https://docs.sarvam.ai/llms.txt.
> For full documentation content in one file, see https://docs.sarvam.ai/llms-full.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.sarvam.ai/_mcp/server.

# How to control the response length with `max_tokens`

> control how long the model's response can be

The `max_tokens` parameter lets you control **how long the model's response can be**: in terms of tokens.

* A **token** can be a word, part of a word, or even punctuation\
  (Example: "Hello!" = 2 tokens: `"Hello"` + `"!"`)

### Why use `max_tokens`?

* To **limit the size** of the output
* To **control latency / cost** (fewer tokens = faster and cheaper)
* To avoid overly long answers if you want concise responses

---

### How to choose the value:

### Parameter details:

| Parameter    | Type    | Default |
| ------------ | ------- | ------- |
| `max_tokens` | Integer | 2048    |

---

First, install the SDK:

```bash
pip install -Uqq sarvamai
```

Then use the following Python code:

```python
from sarvamai import SarvamAI

# Initialize the SarvamAI client with your API key
client = SarvamAI(api_subscription_key="YOUR_SARVAM_API_KEY")

# Example 1: Using default max_tokens (not specified): defaults to 2048
response = client.chat.completions(
    model="sarvam-105b",
    messages=[
        {"role": "system", "content": "You are a helpful assistant."},
        {"role": "user", "content": "Tell me about the planet Mars."}
    ],
    # max_tokens not specified → defaults to 2048
)

print(response.choices[0].message.content)
```

```python
from sarvamai import SarvamAI

client = SarvamAI(api_subscription_key="YOUR_SARVAM_API_KEY")

# Example 2: Using max_tokens = 100: limit response length
response = client.chat.completions(
    model="sarvam-105b",
    messages=[
        {"role": "system", "content": "You are a helpful assistant."},
        {"role": "user", "content": "Summarize the plot of Mahabharata."}
    ],
    max_tokens=100
)

# Receive assistant's reply as output
print(response.choices[0].message.content)
```

> control how long the model's response can be