# Hugging Face — DABStep: Data Agent Benchmark for Multi-step Reasoning

- Company: Hugging Face (huggingface.co)
- Announced: 2025-02-04T00:00:00+00:00
- Category: research-paper
- Subject: Platform
- Source: https://huggingface.co/blog/dabstep
- Record: https://forck.live/items/1755-dabstep-data-agent-benchmark-for-multi-step-reasoning

Hugging Face and Adyen introduce DABStep, a benchmark of over 450 real-world data analysis tasks for evaluating LLM agents. Current models achieve only 16% accuracy.

## Evidence

Verbatim from https://huggingface.co/blog/dabstep:

> Adyen and Hugging Face built the Data Agent Benchmark for Multi-step Reasoning (DABstep) together. DABstep consists of over 450 data analysis tasks designed to evaluate the capabilities of state-of-the-art LLMs and AI agents.
> 
> Our findings reveal that DABstep presents a significant challenge for current AI models, with the most capable reasoning-based agents achieving only 16% accuracy, highlighting significant progress to be made in the field.

---

Record: https://forck.live/items/1755-dabstep-data-agent-benchmark-for-multi-step-reasoning
Catalogue: https://forck.live/llms.txt
Feed: https://forck.live/feed.md