~/thinkingdbx/blog/intelligence-economics
← all articles
thinkingdbx / blog
notes from the notebook
case study
Article #15 June 11, 2026 5 min read AI / Economics NEW

Intelligence Is Deflating:
938× Cheaper in Three Years

MM
Mallesh Madapathi
Founder & CEO, ThinkingDBx

GPT-4-class intelligence got 938× cheaper in three years. Intelligence is deflating faster than any technology input in history.

I built a pipeline on Bonacci Studio that snapshots live AI pricing, capability benchmarks, and GPU rental costs, then asks a simple question: if you hold capability fixed, what does it cost over time?

Hold Capability Fixed

Hold it at GPT-4's intelligence — its 2023 benchmark score — and track the cheapest model that still clears that bar:

Mar 2023 GPT-4 $37.50 / M tokens
May 2024 GPT-4o $7.50
Sep 2024 Qwen2.5 72B $0.37
Aug 2025 gpt-oss-20B $0.088
Mar 2026 Qwen3.5 2B $0.04

The capability that defined the frontier in 2023 is now served by a 2-billion-parameter model at four cents per million tokens.

938×
cost reduction in 3 years
13×
price spread, same open model
54×
API cheaper than self-hosting

Three Things the Data Made Sharper

The Pipeline That Keeps Running

The interesting part isn't the snapshot — it's that the pipeline keeps running. Scheduled ingestion, change detection that flags the moment a provider cuts a price, an append-only store that becomes a proprietary dataset the longer it runs. History you can't buy after the fact.

Intelligence is deflating ~10× a year at constant capability. The question for anyone building on it isn't "is it getting cheaper" — it's "am I capturing the deflation, or pricing as if it stopped."

What you're seeing below: the full pricing pipeline walkthrough — live model snapshot ingestion, capability crosswalk, cost-per-unit analysis, and provider spread charts. Built and run entirely in Bonacci Studio.
studio.bonacci.thinkingdbx.com — Economics of Intelligence
#AI #LLM #DataEngineering #BonacciStudio #MLOps #thinkingdbx

— more soon ✎