Implementation:Openai Openai python Input Token Count Params

Knowledge Sources	Openai_Openai_python OpenAI API Reference
Domains	API_Types, Responses_API
Last Updated	2026-02-15 00:00 GMT

Overview

Concrete parameter type for counting input tokens before submitting a Responses API request, provided by the openai-python SDK.

Description

InputTokenCountParams is a TypedDict that mirrors the shape of a Responses API create request for the purpose of token counting. It includes all the major request fields such as conversation, input, instructions, model, reasoning, text, tools, tool_choice, truncation, and others. The type also defines helper type aliases Conversation (a union of str and ResponseConversationParam), Text (a TypedDict with format and verbosity), and ToolChoice (a union of multiple tool choice param types). This type is auto-generated from the OpenAI OpenAPI specification by Stainless.

Usage

Import InputTokenCountParams when you need to estimate input token counts for a planned Responses API request without actually executing it.

Code Reference

Source Location

• Repository: openai-python
• File: src/openai/types/responses/input_token_count_params.py

Signature

class InputTokenCountParams(TypedDict, total=False):
    conversation: Optional[Conversation]
    input: Union[str, Iterable[ResponseInputItemParam], None]
    instructions: Optional[str]
    model: Optional[str]
    parallel_tool_calls: Optional[bool]
    previous_response_id: Optional[str]
    reasoning: Optional[Reasoning]
    text: Optional[Text]
    tool_choice: Optional[ToolChoice]
    tools: Optional[Iterable[ToolParam]]
    truncation: Literal["auto", "disabled"]

Import

from openai.types.responses import InputTokenCountParams

I/O Contract

Fields

Name	Type	Required	Description
conversation	Optional[Union[str, ResponseConversationParam]]	No	The conversation this response belongs to. Items from this conversation are prepended to input items.
input	Union[str, Iterable[ResponseInputItemParam], None]	No	Text, image, or file inputs to the model.
instructions	Optional[str]	No	A system (or developer) message inserted into the model context.
model	Optional[str]	No	Model ID used to generate the response, like gpt-4o or o3.
parallel_tool_calls	Optional[bool]	No	Whether to allow the model to run tool calls in parallel.
previous_response_id	Optional[str]	No	The unique ID of a previous response for multi-turn conversations.
reasoning	Optional[Reasoning]	No	Configuration options for reasoning models (gpt-5 and o-series only).
text	Optional[Text]	No	Configuration for text response format including format and verbosity.
tool_choice	Optional[ToolChoice]	No	Controls which tool the model should use, if any.
tools	Optional[Iterable[ToolParam]]	No	An array of tools the model may call while generating a response.
truncation	Literal["auto", "disabled"]	No	The truncation strategy for the model response.

Helper Types

Name	Definition	Description
Conversation	Union[str, ResponseConversationParam]	Conversation identifier or configuration.
Text	TypedDict with format and verbosity	Configuration for text response format and verbosity level.
ToolChoice	Union of multiple ToolChoice*Param types	Controls which tool the model should use.

Usage Examples

from openai.types.responses import InputTokenCountParams

params: InputTokenCountParams = {
    "model": "gpt-4o",
    "input": "Explain quantum computing in simple terms.",
    "truncation": "auto",
}

# Use with the input token count endpoint
token_count = client.responses.input_token_count(**params)
print(f"Estimated input tokens: {token_count}")

Related Pages

• Environment:Openai_Openai_python_Python_3_9_Plus