llm4s-docs/org.llm4s/org.llm4s.llmconnect/org.llm4s.llmconnect.model/CompletionOptions

CompletionOptions

org.llm4s.llmconnect.model.CompletionOptions

See theCompletionOptions companion object

@Stable

final case class CompletionOptions

Represents options for a completion request.

Value parameters

budgetTokens: Optional explicit budget for thinking tokens (Anthropic Claude). If set, overrides the default budget from reasoning effort level.
frequencyPenalty: Penalizes new tokens based on their existing frequency in the text so far, discouraging repetition.
maxTokens: Optional maximum number of tokens to generate in the completion.
presencePenalty: Penalizes new tokens based on whether they appear in the text so far, encouraging new topics.
reasoning: Optional reasoning effort level for models that support extended thinking (o1/o3, Claude). For non-reasoning models, this setting is silently ignored.
responseFormat: Optional structured output format (e.g. JSON or JSON schema). Support is provider- and model-dependent; see capability validation in RequestTransformer.
temperature: Controls the randomness of the output. Higher values make the output more random. Note: Reasoning models (o1/o3) ignore this setting.
tools: Optional sequence of tool function definitions that can be requested by the LLM during a completion.
topP: Controls the diversity of the output. Lower values make the output more focused.

Attributes

Example

import org.llm4s.llmconnect.model._
// Enable high reasoning for complex tasks
val options = CompletionOptions()
 .withReasoning(ReasoningEffort.High)
 .copy(maxTokens = Some(4096))
// For Anthropic, set explicit thinking budget
val anthropicOptions = CompletionOptions()
 .withBudgetTokens(16000)
// Request JSON output (provider-dependent)
val jsonOptions = CompletionOptions().withResponseFormat(ResponseFormat.Json)

Companion

object

Graph

Supertypes

trait Serializable

trait Product

trait Equals

class Object

trait Matchable

class Any

Show all

Members list

Value members

Concrete methods

Get the effective budget tokens for Anthropic extended thinking.

Returns explicit budgetTokens if set, otherwise derives from reasoning effort.

Attributes

Check if reasoning is enabled.

Attributes

Sets or clears the explicit budget for thinking tokens (Anthropic Claude). If set, it overrides the default budget from the reasoning effort level.

Value parameters

budgetTokens: the thinking token budget; None sets no explicit budget (the default), so the budget comes from the reasoning effort level, if any

Attributes

Returns: new CompletionOptions with budgetTokens set

Enable reasoning with explicit token budget for thinking (Anthropic Claude).

This overrides the default budget from the reasoning effort level.

Value parameters

tokens: the number of tokens to budget for thinking

Attributes

Returns: new CompletionOptions with budget tokens set

Sets the frequency penalty, which penalizes new tokens based on their existing frequency in the text so far, discouraging repetition.

Value parameters

frequencyPenalty: the frequency penalty (default 0.0)

Attributes

Returns: new CompletionOptions with frequencyPenalty set

Limits the number of tokens generated in the completion.

Value parameters

maxTokens: the maximum number of tokens to generate

Attributes

Returns: new CompletionOptions with maxTokens set to Some(maxTokens)

Sets or clears the limit on the number of tokens generated in the completion.

Value parameters

maxTokens: the maximum number of tokens to generate; None sets no limit (the default)

Attributes

Returns: new CompletionOptions with maxTokens set

Sets the presence penalty, which penalizes new tokens based on whether they appear in the text so far, encouraging new topics.

Value parameters

presencePenalty: the presence penalty (default 0.0)

Attributes

Returns: new CompletionOptions with presencePenalty set

Sets or clears the reasoning effort level for models that support extended thinking (o1/o3, Claude). For non-reasoning models, this setting is silently ignored.

Value parameters

reasoning: the reasoning effort level; None sets no reasoning effort (the default)

Attributes

Returns: new CompletionOptions with reasoning set

Enable reasoning with the specified effort level.

Value parameters

effort: the reasoning effort level

Attributes

Returns: new CompletionOptions with reasoning enabled

Sets or clears the structured output format (e.g. JSON or JSON schema). Support is provider- and model-dependent; see capability validation in RequestTransformer.

Value parameters

responseFormat: the desired response format; None requests no particular format (the default)

Attributes

Returns: new CompletionOptions with responseFormat set

Request structured output in the given format.

Support is provider- and model-dependent. Capability validation is applied via RequestTransformer (Json: fallback when unsupported; JsonSchema: strict error).

Value parameters

format: the desired response format (e.g. ResponseFormat.Json or ResponseFormat.JsonSchema(schema))

Attributes

Returns: new CompletionOptions with responseFormat set

Sets the sampling temperature. Higher values make the output more random. Reasoning models (o1/o3) ignore this setting.

Value parameters

temperature: the sampling temperature (default 0.7)

Attributes

Returns: new CompletionOptions with temperature set

Sets the tool function definitions that the LLM can request during a completion. Replaces any tools set before.

Value parameters

tools: the tools to offer; an empty sequence offers none (the default)

Attributes

Returns: new CompletionOptions with tools set

Sets the nucleus sampling value. Lower values make the output more focused.

Value parameters

topP: the top-p value (default 1.0)

Attributes

Returns: new CompletionOptions with topP set

Inherited methods

Attributes

Inherited from:: Product

Attributes

Inherited from:: Product

In this article

Generated with