Property ReasoningLevel

Namespace

LMKit.Agents.Orchestration

Assembly

LM-Kit.NET.dll

ReasoningLevel

Gets or sets the model-internal reasoning level applied to every agent execution within this orchestration.

Propagated to ReasoningLevel. null (default) preserves each agent's underlying conversation reasoning level (ReasoningLevel, by default the model's own: Medium, or None for a model whose chat template reasons only when asked). Set this to None when running orchestrations against thinking-capable models where chain-of-thought emission is unwanted (e.g. fast classification pipelines).

public ReasoningLevel? ReasoningLevel { get; set; }

Property Value

ReasoningLevel?

Examples

Disabling reasoning across an entire orchestration (supervisor + workers):

using LMKit.Model;
using LMKit.Agents;
using LMKit.Agents.Orchestration;
using LMKit.TextGeneration.Chat;
using var model = LM.LoadFromModelID("qwen3:8b");
var supervisor = Agent.CreateBuilder(model).WithPersona("Supervisor").Build();
var writer = Agent.CreateBuilder(model).WithPersona("Writer").Build();
var orchestrator = new SupervisorOrchestrator(supervisor)
.AddWorker("writer", writer);
// Both supervisor and worker run without <think> emission for this call.
var result = await orchestrator.ExecuteAsync(
"Write a tagline for a coffee brand.",
new OrchestrationOptions
{
ReasoningLevel = ReasoningLevel.None
});