This AI Agent Picks Its Own Brain (10x Cheaper, n8n)

This AI Agent Picks Its Own Brain (10x Cheaper, n8n)

by Nate Herk

Applied Sciences & Engineering AI — Models & Research UYQArtificial intelligence (AI)UYQMMachine learning
🎙 Nate Herk 👥 964K 📅 April 30, 2025 ⏱ 13 min 👁 41K 📄 tutorial 🧭 2026-08-28
Available in: English (current) Français

Keywords

dynamic model selectionn8nOpenRoutercost optimizationAI agent

Summary

This video presents a no-code system built in n8n that allows an AI agent to dynamically select the most appropriate language model for each task, aiming to reduce costs and improve performance. The creator demonstrates a live demo where the agent, interacting via Slack, chooses between models like Gemini 2.0 Flash, GPT-4.1 Mini, Claude 3.7 Sonnet, and OpenAI’s o1 reasoning model based on the complexity of the request. The system uses a first agent (model selector) with a system prompt defining model strengths, and a second agent (dynamic brain) that uses the selected model to execute the task. The video also covers logging outputs to Google Sheets, comparing models using tools like Vellum and LMArena, and shows a RAG agent example with Supabase. The creator emphasizes the benefits of custom model selection over OpenRouter’s auto mode, and provides resources for downloading the workflow.

143 words

Critical Evaluation

Value of the Information & Strength of the Argument

The video offers practical value by demonstrating a concrete implementation of dynamic model routing, which is a relevant topic for AI practitioners looking to optimize costs. The argumentation is based on live demonstrations and real-world examples, showing the system’s effectiveness in different scenarios. However, the claims of ‘10x cheaper’ are not rigorously quantified, and the selection criteria are based on subjective model strengths rather than empirical benchmarks. The creator’s reasoning is clear and logical, but the lack of comparative cost analysis weakens the overall argument.

Scientific Rigor, Source Quality, Title Accuracy

The video cites several tools and resources, including OpenRouter, Vellum’s LLM leaderboard, and LMArena, which are relevant for model comparison. However, these are not academic sources, and the creator does not provide any formal references or data to support the claims. The title accurately reflects the content, and the video is well-structured with clear timestamps. The promotional nature of the content (e.g., affiliate links, community promotion) is transparent but may introduce bias. No comments were provided for analysis.

178 words

Title / Content Match

The title accurately reflects the content: the video demonstrates an AI agent that dynamically selects its own language model, with claims of cost reduction and performance optimization.

Quality & Reliability

7/10

The video provides a clear, practical tutorial on building a dynamic model selection system in n8n, with live demonstrations and concrete examples. The approach is reproducible and based on real tools (OpenRouter, n8n, Slack, Supabase). However, the content is largely promotional, with limited critical analysis of model selection criteria or potential biases. The sources cited are mostly tool links and community resources, not academic or technical references.

Chapters

Cited Sources

Concurring Sources

  • OpenRouter — The platform used for model routing, which supports the feasibility of the approach.

Contribution & Novelties

The video’s original contribution is a practical, no-code implementation of dynamic model selection within an agent workflow, which is a relatively novel approach for non-programmers. It demonstrates how to leverage OpenRouter’s API to route tasks to different models based on a system prompt, offering a customizable alternative to automatic routing. The inclusion of a RAG example further illustrates the versatility of the approach.

Pour aller plus loin :

  • OpenRouter documentation — Official documentation for the API used, providing details on model routing and parameters.
  • n8n documentation — Official n8n documentation for building workflows, including HTTP requests and AI agent nodes.
  • LLM Routing: A Survey — Note: This is a placeholder; actual survey on LLM routing may exist, but without a verified URL, it’s safer to cite the concept without a link. — For academic background on model selection and routing strategies.

141 words

Radar Profile

The radar profile shows balanced scores across all dimensions, with slightly lower technical depth and information quantity compared to quality and reliability. This indicates a practical, well-executed tutorial that is accessible but not deeply technical.

Reliability 7/10