BesiegeField - Car - Iterative Editing: leaderboard
Metric: Mean distance (m). Source: besiegefield.github.io. Saturation forecast: Around November 2026. 8 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 2.5 Pro | 34.34 |
| 2 | Qwen 3 Coder 480B A35B | 15.24 |
| 3 | O3 | 14.34 |
| 4 | DeepSeek V3 | 8.23 |
| 5 | Claude Opus 4 (20250514) | 8.07 |
| 6 | Llama 4 Scout | 3.04 |
Interactive version: theaggregate.ai/benchmark?slug=besiegefield-car-iterative-editing · How It Works · Data refreshed daily, snapshot 2026-10-09.