OpenAIAssistantInvocationOptions Class
Definition
Important
Some information relates to prerelease product that may be substantially modified before it’s released. Microsoft makes no warranties, express or implied, with respect to the information provided here.
Defines per invocation execution options that override the assistant definition.
public sealed class OpenAIAssistantInvocationOptionstype OpenAIAssistantInvocationOptions = classPublic NotInheritable Class OpenAIAssistantInvocationOptions- Inheritance
-
OpenAIAssistantInvocationOptions
Remarks
Not applicable to AgentChat usage.
Constructors
| OpenAIAssistantInvocationOptions() |
Properties
| AdditionalInstructions |
Appends additional instructions. |
| EnableCodeInterpreter |
Set if code_interpreter tool is enabled. |
| EnableFileSearch |
Set if file_search tool is enabled. |
| EnableJsonResponse |
Set if json response-format is enabled. |
| MaxCompletionTokens |
The maximum number of completion tokens that may be used over the course of the run. |
| MaxPromptTokens |
The maximum number of prompt tokens that may be used over the course of the run. |
| Metadata |
A set of up to 16 key/value pairs that can be attached to an agent, used for storing additional information about that object in a structured format.Keys may be up to 64 characters in length and values may be up to 512 characters in length. |
| ModelName |
Override the AI model targeted by the agent. |
| ParallelToolCallsEnabled |
Enables parallel function calling during tool use. Enabled by default. Use this property to disable. |
| Temperature |
The sampling temperature to use, between 0 and 2. |
| TopP |
An alternative to sampling with temperature, called nucleus sampling, where the model considers the results of the tokens with top_p probability mass. So 0.1 means only the tokens comprising the top 10% probability mass are considered. |
| TruncationMessageCount |
When set, the thread will be truncated to the N most recent messages in the thread. |