ShapeCodeBench: leaderboard
Metric: Mean foreground IoU (%) between the re-rendered and the target shape pixels on the 150-image eval_v1 split (50 per difficulty tier); the model writes a program in a four-primitive drawing DSL that is re-rendered and compared with the target raster, zero-shot, GPT-5.5 through the Codex CLI and Claude Opus 4.7 through Claude Code at the stated effort; a parse failure scores 0; higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 4 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.5 (xHigh) | 86.5 |
| 2 | GPT-5.5 (Medium) | 85 |
| 3 | Claude Opus 4.7 (Max) | 46.1 |
| 4 | Claude Opus 4.7 (High) | 43.9 |
Interactive version: theaggregate.ai/benchmark?slug=shapecodebench · How It Works · Data refreshed daily, snapshot 2026-10-07.