BoundaryBench (Unrestricted): leaderboard
Metric: Success Rate (%). Source: boundarybench.com. 14 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.6 Sol | 83.9 |
| 2 | Grok 4.5 | 82 |
| 3 | Gemini 3.7 Flash | 79.8 |
| 4 | GPT-5.6 Terra | 79 |
| 5 | Claude Fable 5 | 79 |
| 6 | Claude Opus 5 | 77.5 |
| 7 | Claude Opus 4.8 | 74.9 |
| 8 | GLM-5.2 | 73.4 |
| 9 | Claude Sonnet 5 | 73.4 |
| 10 | Kimi K3 | 72.3 |
| 11 | GPT-5.6 Luna | 71.9 |
| 12 | Muse Spark 1.2 | 70.8 |
| 13 | MiniMax-M3 | 59.6 |
| 14 | Qwen 3.7 Max | 56.9 |
Interactive version: theaggregate.ai/benchmark?slug=boundarybench-unrestricted · How It Works · Data refreshed daily, snapshot 2026-09-21.