---
format: "aidr-story-markdown/v1"
id: "af231b9b865928cdc34e4952920008fd1d817beb3f3e036ebf8cf4f9df889612"
canonical_url: "https://aidr.today/af231b9b?lang=en"
title: "VibePrevisBench: Evaluating Agentic Coding for 4D Scene Generation in Blender"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-11-16T00:00:00.000Z"
category: "Data"
topics: ["agent","data-engineering"]
source_urls: ["https://huggingface.co/papers/2611.15789"]
summary: "VibePrevisBench is a new benchmark for evaluating agentic coding for 4D scene reconstruction in Blender. The benchmark consists of 242 tasks where an agent turns a brief into an executable 4D scene and renders a proxy video. Evaluation is conducted through a rule-based Scene Check and a VLM-based Screen Check."
---

# VibePrevisBench: Evaluating Agentic Coding for 4D Scene Generation in Blender

> [Open the canonical story](<https://aidr.today/af231b9b?lang=en>)

**Published:** 2026-11-16T00:00:00.000Z
**Category:** Data
**Topics:** agent, data\-engineering

## Summary

VibePrevisBench is a new benchmark for evaluating agentic coding for 4D scene reconstruction in Blender\. The benchmark consists of 242 tasks where an agent turns a brief into an executable 4D scene and renders a proxy video\. Evaluation is conducted through a rule\-based Scene Check and a VLM\-based Screen Check\.

## Sources

- [Story source](<https://huggingface.co/papers/2611.15789>)

