R2ABench (Mini-SWE-Agent) - GED Accuracy: leaderboard
Metric: Normalized graph-edit-distance accuracy (0-100) between the generated and reference architecture graphs (L1), over the 68 R2ABench projects (17 educational and 51 GitHub repositories), generating a PlantUML architecture view from the structured requirements specification with the Mini-SWE-Agent framework under default settings and greedy decoding; scored only over the outputs that render and parse (the L0 gate), failed outputs excluded; higher is better. Source: arxiv.org. Saturation forecast: Around December 2027. 4 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3 Coder 480B A35B | 51.76 |
| 2 | GPT-5 | 49.31 |
| 3 | Claude Sonnet 4.6 | 48.84 |
| 4 | DeepSeek V3.2 | 45.93 |
Interactive version: theaggregate.ai/benchmark?slug=r2abench-mini-swe-agent-ged-accuracy · How It Works · Data refreshed daily, snapshot 2026-10-07.