OpenAI APIs - Completions

SGLang provides OpenAI-compatible APIs to enable a smooth transition from OpenAI services to self-hosted local models. A complete reference for the API is available in the OpenAI API Reference.

This tutorial covers the following popular APIs:

• chat/completions
• completions

Check out other tutorials to learn about vision APIs for vision-language models and embedding APIs for embedding models.

Launch A Server

Launch the server in your terminal and wait for it to initialize.

Example

from sglang.test.doc_patch import launch_server_cmd
from sglang.utils import wait_for_server, print_highlight, terminate_process

server_process, port = launch_server_cmd(
    "python3 -m sglang.launch_server --model-path qwen/qwen2.5-0.5b-instruct --host 0.0.0.0 --log-level warning"
)

wait_for_server(f"http://localhost:{port}")
print(f"Server started on http://localhost:{port}")

Chat Completions

Usage

The server fully implements the OpenAI API. It will automatically apply the chat template specified in the Hugging Face tokenizer, if one is available. You can also specify a custom chat template with --chat-template when launching the server.

Example

import openai

client = openai.Client(base_url=f"http://127.0.0.1:{port}/v1", api_key="None")

response = client.chat.completions.create(
    model="qwen/qwen2.5-0.5b-instruct",
    messages=[
        {"role": "user", "content": "List 3 countries and their capitals."},
    ],
    temperature=0,
    max_tokens=64,
)

print_highlight(f"Response: {response}")

Model Thinking/Reasoning Support

Some models support internal reasoning or thinking processes that can be exposed in the API response. SGLang provides unified support for various reasoning models through the chat_template_kwargs parameter and compatible reasoning parsers.

Supported Models and Configuration

Model Family	Chat Template Parameter	Reasoning Parser	Notes
DeepSeek-R1 (R1, R1-0528, R1-Distill)	enable_thinking	--reasoning-parser deepseek-r1	Standard reasoning models
DeepSeek-V3.1	thinking	--reasoning-parser deepseek-v3	Hybrid model (thinking/non-thinking modes)
Qwen3 (standard)	enable_thinking	--reasoning-parser qwen3	Hybrid model (thinking/non-thinking modes)
Qwen3-Thinking	N/A (always enabled)	--reasoning-parser qwen3-thinking	Always generates reasoning
Kimi	N/A (always enabled)	--reasoning-parser kimi	Kimi thinking models
Gpt-Oss	N/A (always enabled)	--reasoning-parser gpt-oss	Gpt-Oss thinking models

Basic Usage

To use reasoning output:

1. Launch the server with the appropriate reasoning parser.
2. Use the model-specific parameter in chat_template_kwargs when you need to override the model’s chat-template default.
3. Optionally use separate_reasoning: False to keep reasoning in the response content instead of returning it separately (defaults to True).

Note for Qwen3-Thinking models: These models always generate thinking content and do not support the enable_thinking parameter. Use --reasoning-parser qwen3-thinking or --reasoning-parser qwen3 to parse the thinking content.

Example: Qwen3 Models

Example

# Launch server:
# python3 -m sglang.launch_server --model Qwen/Qwen3-4B --reasoning-parser qwen3

from openai import OpenAI

client = OpenAI(
    api_key="EMPTY",
    base_url=f"http://127.0.0.1:30000/v1",
)

model = "Qwen/Qwen3-4B"
messages = [{"role": "user", "content": "How many r's are in 'strawberry'?"}]

response = client.chat.completions.create(
    model=model,
    messages=messages,
    extra_body={
        "chat_template_kwargs": {"enable_thinking": True},
        "separate_reasoning": True
    }
)

print("Reasoning:", response.choices[0].message.reasoning_content)
print("-"*100)
print("Answer:", response.choices[0].message.content)

ExampleOutput:

Output

Reasoning: Okay, so the user is asking how many 'r's are in the word 'strawberry'. Let me think. First, I need to make sure I have the word spelled correctly. Strawberry... S-T-R-A-W-B-E-R-R-Y. Wait, is that right? Let me break it down.

Starting with 'strawberry', let's write out the letters one by one. S, T, R, A, W, B, E, R, R, Y. Hmm, wait, that's 10 letters. Let me check again. S (1), T (2), R (3), A (4), W (5), B (6), E (7), R (8), R (9), Y (10). So the letters are S-T-R-A-W-B-E-R-R-Y.
...
Therefore, the answer should be three R's in 'strawberry'. But I need to make sure I'm not counting any other letters as R. Let me check again. S, T, R, A, W, B, E, R, R, Y. No other R's. So three in total. Yeah, that seems right.

----------------------------------------------------------------------------------------------------
Answer: The word "strawberry" contains **three** letters 'r'. Here's the breakdown:

1. **S-T-R-A-W-B-E-R-R-Y**
   - The **third letter** is 'R'.
   - The **eighth and ninth letters** are also 'R's.

Thus, the total count is **3**.

**Answer:** 3.

Setting "enable_thinking": False disables reasoning for Qwen3. Omitting the parameter follows the model’s chat-template default; Qwen3 enables thinking by default. Qwen3-Thinking models always generate reasoning content and do not support the enable_thinking parameter.

Set a Default for OpenAI Chat Completions

If most OpenAI chat-completion requests should use the same chat-template mode, set a default when launching the server:

Command

python -m sglang.launch_server \
  --model-path Qwen/Qwen3-4B \
  --reasoning-parser qwen3 \
  --default-chat-template-kwargs '{"enable_thinking": false}'

--default-chat-template-kwargs accepts a JSON object and merges it into each OpenAI chat-completion request’s chat_template_kwargs. For ordinary template keys such as enable_thinking, the precedence is:

1. An explicit value in the request’s chat_template_kwargs
2. The value in --default-chat-template-kwargs
3. The model’s chat-template default

reasoning_effort follows a separate normalization path, so the precedence list above does not apply to it. Configure reasoning_effort in either the server defaults or individual requests, not both.

The accepted keys are model-specific. For example, Qwen3 uses enable_thinking, while other model families may use thinking or may not expose a reasoning toggle. The reasoning parser controls how reasoning tokens are separated from the response; it does not by itself enable reasoning in the chat template.

Logit Bias Support

SGLang supports the logit_bias parameter for both chat completions and completions APIs. This parameter allows you to modify the likelihood of specific tokens being generated by adding bias values to their logits. The bias values can range from -100 to 100, where:

• Positive values (0 to 100) increase the likelihood of the token being selected
• Negative values (-100 to 0) decrease the likelihood of the token being selected
• -100 effectively prevents the token from being generated

The logit_bias parameter accepts a dictionary where keys are token IDs (as strings) and values are the bias amounts (as floats).

Getting Token IDs

To use logit_bias effectively, you need to know the token IDs for the words you want to bias. Here’s how to get token IDs:

Example

# Get tokenizer to find token IDs
import tiktoken

# For OpenAI models, use the appropriate encoding
tokenizer = tiktoken.encoding_for_model("gpt-3.5-turbo")  # or your model

# Get token IDs for specific words
word = "sunny"
token_ids = tokenizer.encode(word)
print(f"Token IDs for '{word}': {token_ids}")

# For SGLang models, you can access the tokenizer through the client
# and get token IDs for bias

Important: The logit_bias parameter uses token IDs as string keys, not the actual words.

Example: DeepSeek-V3 Models

DeepSeek-V3 models support thinking mode through the thinking parameter:

Example

# Launch server:
# python3 -m sglang.launch_server --model deepseek-ai/DeepSeek-V3.1 --tp 8  --reasoning-parser deepseek-v3

from openai import OpenAI

client = OpenAI(
    api_key="EMPTY",
    base_url=f"http://127.0.0.1:30000/v1",
)

model = "deepseek-ai/DeepSeek-V3.1"
messages = [{"role": "user", "content": "How many r's are in 'strawberry'?"}]

response = client.chat.completions.create(
    model=model,
    messages=messages,
    extra_body={
        "chat_template_kwargs": {"thinking": True},
        "separate_reasoning": True
    }
)

print("Reasoning:", response.choices[0].message.reasoning_content)
print("-"*100)
print("Answer:", response.choices[0].message.content)

Example Output:

Output

Reasoning: First, the question is: "How many r's are in 'strawberry'?"

I need to count the number of times the letter 'r' appears in the word "strawberry".

Let me write out the word: S-T-R-A-W-B-E-R-R-Y.

Now, I'll go through each letter and count the 'r's.
...
So, I have three 'r's in "strawberry".

I should double-check. The word is spelled S-T-R-A-W-B-E-R-R-Y. The letters are at positions: 3, 8, and 9 are 'r's. Yes, that's correct.

Therefore, the answer should be 3.
----------------------------------------------------------------------------------------------------
Answer: The word "strawberry" contains **3** instances of the letter "r". Here's a breakdown for clarity:

- The word is spelled: S-T-R-A-W-B-E-R-R-Y
- The "r" appears at the 3rd, 8th, and 9th positions.

DeepSeek-V3 models use the thinking parameter (not enable_thinking) to control reasoning output.

Example

# Example with logit_bias parameter
# Note: You need to get the actual token IDs from your tokenizer
# For demonstration, we'll use some example token IDs
response = client.chat.completions.create(
    model="qwen/qwen2.5-0.5b-instruct",
    messages=[
        {"role": "user", "content": "Complete this sentence: The weather today is"}
    ],
    temperature=0.7,
    max_tokens=20,
    logit_bias={
        "12345": 50,  # Increase likelihood of token ID 12345
        "67890": -50,  # Decrease likelihood of token ID 67890
        "11111": 25,  # Slightly increase likelihood of token ID 11111
    },
)

print_highlight(f"Response with logit bias: {response.choices[0].message.content}")

Parameters

The chat completions API accepts OpenAI Chat Completions API’s parameters. Refer to OpenAI Chat Completions API for more details.

SGLang extends the standard API with the extra_body parameter, allowing for additional customization. One key option within extra_body is chat_template_kwargs, which can be used to pass arguments to the chat template processor.

Example

response = client.chat.completions.create(
    model="qwen/qwen2.5-0.5b-instruct",
    messages=[
        {
            "role": "system",
            "content": "You are a knowledgeable historian who provides concise responses.",
        },
        {"role": "user", "content": "Tell me about ancient Rome"},
        {
            "role": "assistant",
            "content": "Ancient Rome was a civilization centered in Italy.",
        },
        {"role": "user", "content": "What were their major achievements?"},
    ],
    temperature=0.3,  # Lower temperature for more focused responses
    max_tokens=128,  # Reasonable length for a concise response
    top_p=0.95,  # Slightly higher for better fluency
    presence_penalty=0.2,  # Mild penalty to avoid repetition
    frequency_penalty=0.2,  # Mild penalty for more natural language
    n=1,  # Single response is usually more stable
    seed=42,  # Keep for reproducibility
)

print_highlight(response.choices[0].message.content)

Streaming mode is also supported.

Logit Bias Support

The completions API also supports the logit_bias parameter with the same functionality as described in the chat completions section above.

Example

stream = client.chat.completions.create(
    model="qwen/qwen2.5-0.5b-instruct",
    messages=[{"role": "user", "content": "Say this is a test"}],
    stream=True,
)
for chunk in stream:
    if chunk.choices[0].delta.content is not None:
        print(chunk.choices[0].delta.content, end="")

Returning Routed Experts (MoE Models)

For MoE models, set return_routed_experts: true in extra_body to return expert routing data. Requires --enable-return-routed-experts server flag. By default, routed_experts is returned in the response-level sglext object as base64-encoded int32 expert IDs in a flattened array with logical shape [num_tokens, num_layers, top_k]. For non-streaming chat responses with return_meta_info: true, routed_experts is returned only in each choice’s meta_info object and omitted from response-level sglext to avoid duplicating the routing data. Other requested extension fields, such as cached_tokens_details and spec_tokens_details, remain in response-level sglext. By default, routed experts cover [0, seqlen - 1), the full available sequence, because RL workflows need them for the full sequence. Set routed_experts_start_len in extra_body to an absolute prefix length to return only [routed_experts_start_len, seqlen - 1). For example, in multi-turn RL rollouts, routed experts for tokens from previous turns have already been collected, so setting this value avoids unnecessary transfer that causes bottlenecks.

Example

# Example with logit_bias parameter for completions API
# Note: You need to get the actual token IDs from your tokenizer
# For demonstration, we'll use some example token IDs
response = client.completions.create(
    model="qwen/qwen2.5-0.5b-instruct",
    prompt="The best programming language for AI is",
    temperature=0.7,
    max_tokens=20,
    logit_bias={
        "12345": 75,  # Strongly favor token ID 12345
        "67890": -100,  # Completely avoid token ID 67890
        "11111": -25,  # Slightly discourage token ID 11111
    },
)

print_highlight(f"Response with logit bias: {response.choices[0].text}")

Completions

Usage

Completions API is similar to Chat Completions API, but without the messages parameter or chat templates.

Example

response = client.completions.create(
    model="qwen/qwen2.5-0.5b-instruct",
    prompt="List 3 countries and their capitals.",
    temperature=0,
    max_tokens=64,
    n=1,
    stop=None,
)

print_highlight(f"Response: {response}")

Parameters

The completions API accepts OpenAI Completions API’s parameters. Refer to OpenAI Completions API for more details.

Here is an example of a detailed completions request:

Example

response = client.completions.create(
    model="qwen/qwen2.5-0.5b-instruct",
    prompt="Write a short story about a space explorer.",
    temperature=0.7,  # Moderate temperature for creative writing
    max_tokens=150,  # Longer response for a story
    top_p=0.9,  # Balanced diversity in word choice
    stop=["\n\n", "THE END"],  # Multiple stop sequences
    presence_penalty=0.3,  # Encourage novel elements
    frequency_penalty=0.3,  # Reduce repetitive phrases
    n=1,  # Generate one completion
    seed=123,  # For reproducible results
)

print_highlight(f"Response: {response}")

Returning Routed Experts (MoE Models)

For MoE models, set return_routed_experts: true in extra_body to return expert routing data. Requires --enable-return-routed-experts server flag. The routed_experts field will be returned in the sgl_ext object on each choice, containing base64-encoded int32 expert IDs as a flattened array with logical shape [num_tokens, num_layers, top_k]. By default this returns [0, seqlen - 1), the full available sequence, because RL workflows need routed experts for the full sequence. Set routed_experts_start_len in extra_body to an absolute prefix length to return only [routed_experts_start_len, seqlen - 1). For example, in multi-turn RL rollouts, routed experts for tokens from previous turns have already been collected, so setting this value avoids unnecessary transfer that cause bottlenecks.

Structured Outputs (JSON, Regex, EBNF)

For OpenAI compatible structured outputs API, refer to Structured Outputs for more details.

Using LoRA Adapters

SGLang supports LoRA (Low-Rank Adaptation) adapters with OpenAI-compatible APIs. You can specify which adapter to use directly in the model parameter using the base-model:adapter-name syntax.

Server Setup:

Command

python -m sglang.launch_server \
    --model-path qwen/qwen2.5-0.5b-instruct \
    --enable-lora \
    --lora-paths adapter_a=/path/to/adapter_a adapter_b=/path/to/adapter_b

For more details on LoRA serving configuration, see the LoRA documentation.

API Call:

(Recommended) Use the model:adapter syntax to specify which adapter to use:

Example

response = client.chat.completions.create(
    model="qwen/qwen2.5-0.5b-instruct:adapter_a",  # ← base-model:adapter-name
    messages=[{"role": "user", "content": "Convert to SQL: show all users"}],
    max_tokens=50,
)

Backward Compatible: Using extra_body

The old extra_body method is still supported for backward compatibility:

Example

# Backward compatible method
response = client.chat.completions.create(
    model="qwen/qwen2.5-0.5b-instruct",
    messages=[{"role": "user", "content": "Convert to SQL: show all users"}],
    extra_body={"lora_path": "adapter_a"},  # ← old method
    max_tokens=50,
)

Note: When both model:adapter and extra_body["lora_path"] are specified, the model:adapter syntax takes precedence.

Example

terminate_process(server_process)