Compare LLM Responses Side-by-Side with Google Sheets

This n8n workflow enables seamless side-by-side comparison of responses from two different language models (LLMs) directly in a chat interface. User inputs are duplicated and routed to separate AI Agent nodes, each configured with its own model (e.g., via OpenRouter, Vertex AI, or OpenAI), system prompts, tools, and memory context. Responses are then displayed sequentially in the chat for immediate visual comparison, allowing quick assessment of quality, consistency, and relevance.

All interactions—including user messages, both LLM outputs, and conversation history—are automatically logged to a Google Sheets spreadsheet. This facilitates manual review, scoring, or automated analysis using additional models. The provided template sheet structures data in columns for prompts, responses, timestamps, and metadata, making it ideal for A/B testing during AI agent development.

Key benefits include non-deterministic LLM evaluation without custom coding, time savings in model selection, and scalability for production decisions. Use cases span AI agent prototyping, fine-tuning comparisons across providers, educational analytics for LLM performance tracking, or R&D in natural language processing. Extendable for more models by duplicating logic and updating the sheet.