Tiered Super-Moore’s Law: Price Evolution, Production Frontiers, and Market Competition in Large Language Model Inference Services
Mingdeng Du · 2026
Read Tiered Super-Moore’s Law: Price Evolution, Production Frontiers, and Market Competition in Large Language Model Inference Services by Mingdeng Du (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-28576.txt.
Description
Du gives a first systematic economic study of token prices for large language model inference. The dataset joins OpenRouter data on 318 models, Epoch AI records, and cross-checked milestones from 2020 to 2026. Du finds a roughly 600-fold price drop and proposes a tiered Super-Moore pattern. Economy and mid-tier prices fall faster than Moore's law, while flagship prices show a reasoning premium. A break test points to May 2024 as a shift from tech-driven to competition-driven declines. Software and design gains, not GPU gains, explain most cost cuts.
Catalog record
| CREATOR | Mingdeng Du |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 37375 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.28576