Skip to content
OpenTrain AIFor AI Companies

HFEPX · Eval paper review

Vibe Code Bench: Evaluating AI Models on End-to-End Web Application Development

Hung Tran, Langston Nashold, Rayan Krishnan, Antoine Bigeard +1 more

Published

Mar 4, 2026

Citations

0

Trust level

Moderate

Usefulness score

65/100 (Medium)

Extraction confidence

70% (Moderate)

Derived from extracted protocol signals and abstract evidence.

Rater population

Not reported

Signals refreshed

May 5, 2026

Open arXiv Open DOI Find implementation Shortlist (0) RSS feed

Tags

Automatic Metrics Coding Pairwise Preference Web Browsing

Should you rely on this paper?

This paper has useful evaluation signal, but protocol completeness is partial; pair it with related papers before deciding implementation strategy.

Use this for comparison and orientation, not as your only source.

Best use

Secondary protocol comparison source

Use if you need

A secondary eval reference to pair with stronger protocol papers.

What to verify

Validate the evaluation procedure and quality controls in the full paper before operational use.

Main weakness

No major weakness surfaced.

Human feedback signal
Detected
From extracted signals
Evaluation signal
Detected
Eval setup described
Usefulness for eval research
65/100
Moderate-confidence candidate

Useful as a secondary reference; validate protocol details against neighboring papers.

If you are doing eval pipeline work, start here

Human Eval Papers LLM-as-Judge Papers Pairwise Preference Papers Tool-Use Evaluation Search

Abstract

Code generation has emerged as one of AI's highest-impact use cases, yet existing benchmarks measure isolated tasks rather than the complete "zero-to-one" process of building a working application from scratch. We introduce Vibe Code Bench, a benchmark of 100 web application specifications (50 public validation, 50 held-out test) with 964 browser-based workflows comprising 10,131 substeps, evaluated against deployed applications by an autonomous browser agent. Across 16 frontier models, the best achieves 61.8% accuracy on the test split, revealing that reliable end-to-end application development remains a frontier challenge. We identify self-testing during generation as a strong performance predictor (Pearson r=0.72), and show through a completed human alignment study that evaluator selection materially affects outcomes (31.8-93.6% pairwise step-level agreement). Our contributions include (1) a novel benchmark dataset and browser-based evaluation pipeline for end-to-end web application development, (2) a comprehensive evaluation of 16 frontier models with cost, latency, and error analysis, and (3) an evaluator alignment protocol with both cross-model and human annotation results.

What we could verify

These are the protocol signals we could actually recover from the available paper metadata. Use them to decide whether this paper is worth deeper reading.

Human Feedback Types

strong

Pairwise Preference

Directly usable for protocol triage.

"Code generation has emerged as one of AI's highest-impact use cases, yet existing benchmarks measure isolated tasks rather than the complete "zero-to-one" process of building a working application from scratch."

Evaluation Modes

strong

Automatic Metrics

Includes extracted eval setup.

"Code generation has emerged as one of AI's highest-impact use cases, yet existing benchmarks measure isolated tasks rather than the complete "zero-to-one" process of building a working application from scratch."

Quality Controls

missing

Not reported

No explicit QC controls found.

"Code generation has emerged as one of AI's highest-impact use cases, yet existing benchmarks measure isolated tasks rather than the complete "zero-to-one" process of building a working application from scratch."

Benchmarks / Datasets

missing

Not extracted

No benchmark anchors detected.

"Code generation has emerged as one of AI's highest-impact use cases, yet existing benchmarks measure isolated tasks rather than the complete "zero-to-one" process of building a working application from scratch."

Reported Metrics

strong

Accuracy, Agreement

Useful for evaluation criteria comparison.

"Across 16 frontier models, the best achieves 61.8% accuracy on the test split, revealing that reliable end-to-end application development remains a frontier challenge."

Benchmarks and datasets

No benchmark or dataset names were extracted from the available abstract.

Reported metrics

accuracyagreement
Human feedback details
Uses human feedback
Yes
Feedback types
Pairwise Preference
Rater population
Not reported
Unit of annotation
Pairwise
Expertise required
Coding
Evaluation details
Evaluation modes
Automatic Metrics
Agentic eval
Web Browsing
Quality controls
Not reported
Evidence quality
Moderate
Use this page as
Secondary protocol comparison source

Research brief

Metadata summary

Code generation has emerged as one of AI's highest-impact use cases, yet existing benchmarks measure isolated tasks rather than the complete "zero-to-one" process of building a working application from scratch.

Based on abstract + metadata only. Check the source paper before making high-confidence protocol decisions.

Key takeaways

  • Code generation has emerged as one of AI's highest-impact use cases, yet existing benchmarks measure isolated tasks rather than the complete "zero-to-one" process of building a working application from scratch.
  • We introduce Vibe Code Bench, a benchmark of 100 web application specifications (50 public validation, 50 held-out test) with 964 browser-based workflows comprising 10,131 substeps, evaluated against deployed applications by an autonomous browser agent.
  • Across 16 frontier models, the best achieves 61.8% accuracy on the test split, revealing that reliable end-to-end application development remains a frontier challenge.

Researcher actions

  • Compare this paper against nearby papers in the same arXiv category before using it for protocol decisions.
  • Validate inferred eval signals (Automatic metrics) against the full paper.
  • Use related-paper links to find stronger protocol-specific references.

Caveats

  • Generated from abstract + metadata only; no PDF parsing.
  • Signals below are heuristic and may miss details reported outside the abstract.

Recommended queries

Pairwise preference evaluation Accuracy

Contribution summary

  • Code generation has emerged as one of AI's highest-impact use cases, yet existing benchmarks measure isolated tasks rather than the complete "zero-to-one" process of building a working application from scratch.
  • We introduce Vibe Code Bench, a benchmark of 100 web application specifications (50 public validation, 50 held-out test) with 964 browser-based workflows comprising 10,131 substeps, evaluated against deployed applications by an autonomous…
  • We identify self-testing during generation as a strong performance predictor (Pearson r=0.72), and show through a completed human alignment study that evaluator selection materially affects outcomes (31.8-93.6% pairwise step-level…

Why it matters for eval

  • We introduce Vibe Code Bench, a benchmark of 100 web application specifications (50 public validation, 50 held-out test) with 964 browser-based workflows comprising 10,131 substeps, evaluated against deployed applications by an autonomous…
  • We identify self-testing during generation as a strong performance predictor (Pearson r=0.72), and show through a completed human alignment study that evaluator selection materially affects outcomes (31.8-93.6% pairwise step-level…

Researcher checklist

  • Human feedback protocol is explicit

    Detected: Pairwise Preference

  • Evaluation mode is explicit

    Detected: Automatic Metrics

  • Quality control reporting appears

    No calibration/adjudication/IAA control explicitly detected.

  • Benchmark or dataset anchors are present

    No benchmark/dataset anchor extracted from abstract.

  • Metric reporting is present

    Detected: accuracy, agreement