This compatibility layer is primarily intended to test and compare model capabilities, and is not considered a long-term or production-ready solution for most use cases. While we do intend to keep it fully functional and not make breaking changes, our priority is the reliability and effectiveness of the Claude API.For more information on known compatibility limitations, see Important OpenAI compatibility limitations.If you encounter any issues with the OpenAI SDK compatibility feature, please let us know here.
For the best experience and access to Claude API full feature set (PDF processing, citations, extended thinking, and prompt caching), we recommend using the native Claude API.

​
Getting started with the OpenAI SDK

To use the OpenAI SDK compatibility feature, you’ll need to:
  1. Use an official OpenAI SDK
  2. Change the following
    • Update your base URL to point to the Claude API
    • Replace your API key with an Claude API key
    • Update your model name to use a Claude model
  3. Review the documentation below for what features are supported

​
Quick start example

from openai import OpenAI

client = OpenAI(
    api_key="ANTHROPIC_API_KEY",  # Your Claude API key
    base_url="https://api.anthropic.com/v1/"  # the Claude API endpoint
)

response = client.chat.completions.create(
    model="claude-opus-4-1-20250805", # Anthropic model name
    messages=[
        {"role": "system", "content": "You are a helpful assistant."},
        {"role": "user", "content": "Who are you?"}
    ],
)

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

​
Important OpenAI compatibility limitations

​
API behavior

Here are the most substantial differences from using OpenAI:
  • The strict parameter for function calling is ignored, which means the tool use JSON is not guaranteed to follow the supplied schema.
  • Audio input is not supported; it will simply be ignored and stripped from input
  • Prompt caching is not supported, but it is supported in the Anthropic SDK
  • System/developer messages are hoisted and concatenated to the beginning of the conversation, as Anthropic only supports a single initial system message.
Most unsupported fields are silently ignored rather than producing errors. These are all documented below.

​
Output quality considerations

If you’ve done lots of tweaking to your prompt, it’s likely to be well-tuned to OpenAI specifically. Consider using our prompt improver in the Claude Console as a good starting point.

​
System / Developer message hoisting

Most of the inputs to the OpenAI SDK clearly map directly to Anthropic’s API parameters, but one distinct difference is the handling of system / developer prompts. These two prompts can be put throughout a chat conversation via OpenAI. Since Anthropic only supports an initial system message, we take all system/developer messages and concatenate them together with a single newline (\n) in between them. This full string is then supplied as a single system message at the start of the messages.

​
Extended thinking support

You can enable extended thinking capabilities by adding the thinking parameter. While this will improve Claude’s reasoning for complex tasks, the OpenAI SDK won’t return Claude’s detailed thought process. For full extended thinking features, including access to Claude’s step-by-step reasoning output, use the native Claude API. 
response = client.chat.completions.create(
    model="claude-opus-4-1-20250805",
    messages=...,
    extra_body={
        "thinking": { "type": "enabled", "budget_tokens": 2000 }
    }
)

​
Rate limits

Rate limits follow Anthropic’s standard limits for the /v1/messages endpoint.

​
Detailed OpenAI Compatible API Support

​
Request fields

​
Simple fields

FieldSupport status
modelUse Claude model names
max_tokensFully supported
max_completion_tokensFully supported
streamFully supported
stream_optionsFully supported
top_pFully supported
parallel_tool_callsFully supported
stopAll non-whitespace stop sequences work
temperatureBetween 0 and 1 (inclusive). Values greater than 1 are capped at 1.
nMust be exactly 1
logprobsIgnored
metadataIgnored
response_formatIgnored
predictionIgnored
presence_penaltyIgnored
frequency_penaltyIgnored
seedIgnored
service_tierIgnored
audioIgnored
logit_biasIgnored
storeIgnored
userIgnored
modalitiesIgnored
top_logprobsIgnored
reasoning_effortIgnored

​
tools / functions fields

tools[n].function fields
FieldSupport status
nameFully supported
descriptionFully supported
parametersFully supported
strictIgnored

​
messages array fields

Fields for messages[n].role == "developer"
Developer messages are hoisted to beginning of conversation as part of the initial system message
FieldSupport status
contentFully supported, but hoisted
nameIgnored

​
Response fields

FieldSupport status
idFully supported
choices[]Will always have a length of 1
choices[].finish_reasonFully supported
choices[].indexFully supported
choices[].message.roleFully supported
choices[].message.contentFully supported
choices[].message.tool_callsFully supported
objectFully supported
createdFully supported
modelFully supported
finish_reasonFully supported
contentFully supported
usage.completion_tokensFully supported
usage.prompt_tokensFully supported
usage.total_tokensFully supported
usage.completion_tokens_detailsAlways empty
usage.prompt_tokens_detailsAlways empty
choices[].message.refusalAlways empty
choices[].message.audioAlways empty
logprobsAlways empty
service_tierAlways empty
system_fingerprintAlways empty

​
Error message compatibility

The compatibility layer maintains consistent error formats with the OpenAI API. However, the detailed error messages will not be equivalent. We recommend only using the error messages for logging and debugging.

​
Header compatibility

While the OpenAI SDK automatically manages headers, here is the complete list of headers supported by the Claude API for developers who need to work with them directly.
HeaderSupport Status
x-ratelimit-limit-requestsFully supported
x-ratelimit-limit-tokensFully supported
x-ratelimit-remaining-requestsFully supported
x-ratelimit-remaining-tokensFully supported
x-ratelimit-reset-requestsFully supported
x-ratelimit-reset-tokensFully supported
retry-afterFully supported
request-idFully supported
openai-versionAlways 2020-10-01
authorizationFully supported
openai-processing-msAlways empty