# BenchCAD Vision2Code voxel IoU with tools (BenchCAD Vision2Code (tools))

> Generates CadQuery code from multi-view renders with image inspection and code-execution tools.

Canonical page: https://benchlm.ai/benchmarks/benchcadvision2codewithtools

- Category: [Multimodal & Grounded](/multimodal-grounded)
- Last updated: September 27, 2026

## About BenchCAD Vision2Code (tools)

- Year: 2026
- Tasks: 1,000-file Vision2Code subset
- Format: Voxel IoU with tools
- Difficulty: Programmatic CAD generation
- Paper: [BenchCAD: A comprehensive, industry-standard benchmark for programmatic CAD](https://arxiv.org/abs/2605.10865)

Section 8.12.2 reports the five-run average on a random 1,000-file subset with max effort, a container, and image cropping.

BenchCAD Vision2Code (tools) is currently displayed on BenchLM for reference, but it is excluded from the weighted scoring formula.

## Leaderboard (3 models)

| Rank | Model | Creator | Score |
|------|-------|---------|-------|
| 1 | [Claude Opus 5.5](/models/claude-opus-5-5) | Anthropic | 0.962 |
| 2 | [GPT-6 Astra](/models/gpt-6-astra) | OpenAI | 0.959 |
| 3 | [Claude Opus 5](/models/claude-opus-5) | Anthropic | 0.821 |

## FAQ

### What does BenchCAD Vision2Code (tools) measure?

Generates CadQuery code from multi-view renders with image inspection and code-execution tools.

### Which model scores highest on BenchCAD Vision2Code (tools)?

Claude Opus 5.5 by Anthropic currently leads with a score of 0.962 on BenchCAD Vision2Code (tools).

### How many models are evaluated on BenchCAD Vision2Code (tools)?

3 AI models have been evaluated on BenchCAD Vision2Code (tools) on BenchLM.

### Does BenchCAD Vision2Code (tools) affect BenchLM's overall score?

Not directly. BenchCAD Vision2Code (tools) is still displayed on BenchLM for reference, but it is excluded from the weighted scoring formula.

## Compare Top Models on BenchCAD Vision2Code (tools)

- [Claude Opus 5.5 vs GPT-6 Astra](/compare/claude-opus-5-5-vs-gpt-6-astra)
- [GPT-6 Astra vs Claude Opus 5](/compare/claude-opus-5-vs-gpt-6-astra)
