Sets or returns the preferred reasoning effort level used for models that support reasoning.
Syntax
object.ReasoningLevel [= value ]
Remarks
The ReasoningLevel property specifies the preferred reasoning
effort level used for requests sent to models which support
reasoning-oriented generation. Higher reasoning levels may improve
response quality for complex prompts, but can also increase response
time and token usage.
| Constant |
Value |
Description |
| llmReasoningNone |
0 |
No reasoning, even if it is available. |
| llmReasoningDefault |
1 |
Provider or model default. |
| llmReasoningLow |
2 |
Prefer faster responses; reduces usage costs. |
| llmReasoningMedium |
3 |
Balanced reasoning. |
| llmReasoningHigh |
4 |
Prefer quality responses; increases usage costs. |
| llmReasoningMaximum |
5 |
Maximum reasoning quality. |
Reasoning support and available reasoning levels depend on the
selected provider and model. Some providers may ignore this property,
map the requested value to a provider-specific setting, or support
only a subset of the available reasoning levels. The default value
is llmReasoningDefault, which allows the provider or
model to select the appropriate reasoning behavior.
This property returns the currently selected reasoning preference.
The returned value does not guarantee that the selected provider or
model supports the requested reasoning level.
Data Type
Integer (Int32)
See Also
InputTokens Property,
MaxTokens Property,
ModelFlags Property,
OutputTokens Property,
Temperature Property,
TotalTokens Property,
SendMessage Method
|