← All Demos

Build vs. Buy: TCO Reality Check

robert@barcik.training

Compare the 12-month total cost of self-hosting open-source models vs. using vendor APIs. Answer 4 simple questions and get a clear cost comparison.

Your Use Case Cost Comparison Recommendation
Your Use Case
Good = small open-source (7–8B) | High = 70B-class / Claude Sonnet | Frontier = Claude Opus / GPT-5.x
12-Month Cost Comparison
Self-Hosted
Vendor API

Self-Hosted Includes

GPU hardware/rental, setup & configuration, monthly electricity & maintenance, ML engineer time for operations, model updates & monitoring.

Vendor API Includes

Per-call pricing (input + output tokens), built-in uptime guarantees, automatic model updates, no infrastructure management needed.

Insights & Recommendation

Based on your inputs, here's what matters for your decision.

Why we stop at 12 months

AI costs are dropping ~30% per year. GPU prices, API pricing, and open-source model quality all shift rapidly. Any projection beyond 12 months is unreliable — plan to re-evaluate in 6–12 months regardless of what you choose today.

Copied to clipboard!
AI transparency

This demo is a scripted, self-contained browser simulation. Nothing you type is sent to an AI model and no live AI system runs behind it, even where it plays one. Its code and copy were built with generative AI (Anthropic’s Claude) and reviewed by Robert Barcik, who is responsible for what is published (LearningDoe s.r.o.). Disclosed in the spirit of Article 50 of the EU AI Act.