---
title: "The Premise of a New S-Curve in AI"
description: "Explore how AI model performance has evolved: smaller models show 32% improvement in 4 months, while Elo scores reveal key advances across model sizes."
categories: ["AI"]
keywords: ["AI model performance","Elo scoring","small AI models","medium AI models","large AI models","mega AI models","algorithm innovations","S-curve in AI"]
ai_summary: "AI model performance has improved significantly, especially in smaller models, indicating a new growth trend in AI development."
date: 2024-10-16
lastmod: 2026-07-21
canonical_url: https://www.tomtunguz.com/elo-improvement/
author: "Tomasz Tunguz"
---

Since [July](https://tomtunguz.com/what-happens-when-model-performance-asymptotes/), have you noticed how much better your AI model has become? Measuring them is hard to do. All we can do is quantify the vibe : is this one better than that one? 

Elo is a score that measures how often one model wins against another, as judged by a human. Which model answers the prompt : "Describe the differences in texture between a Pink Lady and a Macoun apple" better? The one with the higher Elo score.<sup>1</sup>

![image](https://res.cloudinary.com/dzawgnnlr/image/upload/wdq1mlsykpu4yjhdrptr.png)

In the last four months, the top 100 models have improved their Elo by about 60 points, with the top models now at 1339 vs 1287 in July. 



![image](https://res.cloudinary.com/dzawgnnlr/image/upload/uohbmpdtwu6asfvxycdj.png)
The biggest performance gains occurred at the center part of the distribution. Researchers have driven significantly more performance with innovations in algorithms. 

| Model Size | Win Probability Increase (%) | Definition | 
|------------|-----------------------------:| ---:|
| Small      | 32.0%                       | < 10b parameters |
| Medium     | 22.4%                       | 10b - 100b parameters | 
| Large      | 29.6%                       | 100 - 200b parameters |  
| Mega       | 25.9%                       | 200b+ parameters | 

The smallest models have increased performance most. October models have increased their win rates by nearly a third in four months. All of the models have improved their competitive win rates by more than 20%.

![image](https://res.cloudinary.com/dzawgnnlr/image/upload/aleicbp71solzwheqnuu.png)

In July, we posed the question : what happens when model performance asymptotes? Progress in small, medium, & large models is linear in Elo-terms. 

But the mega models show more data points of inflection, suggesting the recent innovations in reasoning & scale (the biggest models have grown from 200b parameters to more than 400b) have produced the beginning of a new high-growth S-curve. 

---
<sup>1</sup> See the [Bradley-Terry model](https://en.wikipedia.org/wiki/Bradley%E2%80%93Terry_model).