---
title: "UC San Diego's AIBuildAI Tops OpenAI's Kaggle Benchmark With 63.1% Medal Rate"
subtitle: "A hierarchical multi-agent system from UC San Diego tops MLE-Bench with a 63.1% medal rate, matching expert Kaggle competitors across 75 tasks."
url: "https://alphasignal.ai/news/uc-san-diego-s-aibuildai-tops-openai-s-kaggle-benchmark-with-63-1-medal-rate"
author: "AlphaSignal Newsroom"
author_url: "https://alphasignal.ai/editorial-team"
company: "AlphaSignal"
company_url: "https://alphasignal.ai/company/alphasignal"
published: "2026-04-15T00:00:00.000Z"
updated: "2026-08-23T20:42:24.833Z"
type: "Paper"
topics:
  - "Agents"
  - "Development"
subtopics:
  - "Code Agents"
  - "Tool Use"
  - "Multi Agent"
  - "Code Agents"
upvotes: 1706
source_url: "https://arxiv.org/abs/2604.14455"
comments: 0
truncated: true
---

# UC San Diego's AIBuildAI Tops OpenAI's Kaggle Benchmark With 63.1% Medal Rate

> A hierarchical multi-agent system from UC San Diego tops MLE-Bench with a 63.1% medal rate, matching expert Kaggle competitors across 75 tasks.

A research team at UC San Diego has released [AIBuildAI](https://github.com/aibuildai/AI-Build-AI), an autonomous agent that takes a task description and a training dataset and hands back a trained model plus an inference script. It just claimed the top spot on [MLE-Bench](https://github.com/openai/mle-bench), OpenAI's benchmark of 75 real Kaggle competitions, with a medal rate high enough to place it above every prior autonomous ML engineering system on a suite curated to test real-world skills like training models, preparing datasets, and running experiments.

The headline number is 63.1%. When MLE-Bench launched, the best setup at the time was OpenAI's o1-preview with AIDE scaffolding, which earned a Kaggle bronze medal in 16.9% of competitions. AIBuildAI nearly quadruples that, and the authors argue it matches the capability of experienced human AI engineers.

## Why AutoML kept hitting a ceiling

Building a competitive ML model is not one job. It is a sequence of very different jobs: sketching a modeling strategy, writing and debugging the training code, watching the loss curves, tweaking hyperparameters, and then ensembling whatever survives. Classic AutoML tools automated a slice of that pipeline, mostly hyperparameter search and model selection inside a predefined space, but they still required humans to define the search space and glue everything together.

The newer wave of LLM-based coding agents like AIDE, AIRA, and MLEvolve reframed the problem as a search over code. Each candidate solution is a script, and a single LLM call mutates one script into a slightly better one. This works, but it caps how much can change in any given step. AIBuildAI's authors argue that a single call cannot hold enough context to redesign an architecture, debug a training loop, and re-tune hyperparameters all at once. Their fix is to stop treating solution refinement as a one-shot code diff.

_Preview only. Full text requires an AlphaSignal Pro account._

## Sources

- [AIBuildAI: An AI Agent for Automatically Building AI Models](https://arxiv.org/html/2604.14455)
- [MLE-bench: Evaluating Machine Learning Agents on Machine Learning Engineering](https://openai.com/index/mle-bench/)
- [MLE-bench arXiv paper](https://arxiv.org/abs/2410.07095)
- [ReX-MLE: Autonomous Agent Benchmark for Medical Imaging Challenges](https://arxiv.org/pdf/2512.17838)
- [AI Research Agents for Machine Learning (AIRA)](https://openreview.net/forum?id=RwfrdKSgCE)