Konnect.ai

Ensemble Mode

Learn how to use Ensemble mode to get synthesized answers from multiple AI models.

What is Ensemble Mode?

Ensemble mode queries multiple AI models simultaneously with your question. Each model provides its own response, and then an aggregation model synthesizes these responses into a single, comprehensive answer that combines the best insights from all models.

Screenshot: Ensemble chat interface

Add image to: /public/docs/ui/ensemble-chat.png

Default models: GPT-5.2 and Claude Sonnet 4.5 are pre-selected when you start a new ensemble chat. You can add or change models before sending your first message.

How to Use Ensemble Mode

1

Select Ensemble Mode

From the dashboard, click on "Ensemble" mode or use the mode selector in the chat input bar.

2

Choose Your Models

Click the "Add" button to select which models to include. You can choose up to 5 models. By default, GPT-5.2 and Claude Sonnet 4.5 are selected.

3

Select Aggregation Method

Choose how responses should be combined. "Synthesis" (default) creates a unified answer from all responses.

4

Ask Your Question

Type your question and press Enter. All selected models will be queried in parallel.

Aggregation Methods

Synthesis

Default

An AI model reads all responses and creates a unified answer that combines the best insights from each model. Best for most use cases.

Best of N

Selects the single best response from all models based on quality and relevance. Best when you want one definitive answer.

Consensus

Highlights points where all models agree, flagging areas of disagreement. Best for fact-checking or understanding confidence levels.

Union

Combines all unique perspectives and information from each model. Best when you want the most comprehensive coverage of a topic.

Understanding the Response

After sending your question, you'll see:

Screenshot: Ensemble response with model tabs

Add image to: /public/docs/ui/ensemble-response.png

Synthesized Response

The main response area shows the aggregated answer combining insights from all models.

Model Tabs

Click on individual model tabs (e.g., "GPT-5.2", "Claude Sonnet 4.5") to see what each model responded individually.

Consensus & Disagreement

A summary section highlights where models agreed and where they differed, helping you understand the confidence level of the response.

Best Practices

  • Mix providers: Include models from different providers (OpenAI, Anthropic, Google) for more diverse perspectives.
  • Complex questions: Ensemble mode shines with nuanced questions that benefit from multiple viewpoints.
  • Follow-up questions: You can ask follow-up questions - all models will receive the synthesized context from the previous response.
  • Check individual responses: If the synthesis seems incomplete, check individual model tabs for additional details.