SAKE - Five-Shot - Software Architecture and Quantum Computing: leaderboard
Metric: Accuracy (%). Source: arxiv.org. 11 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Haiku 4.5 | 95 |
| 2 | GPT-5.4 | 93.8 |
| 3 | Qwen 3 235B A22B | 91.5 |
| 4 | Claude Opus 4.6 | 91.2 |
| 5 | Grok 4.20 | 91.2 |
| 6 | Gemini 3.1 Flash Lite | 90.8 |
| 7 | Mistral Small 4 | 90 |
| 8 | Claude Sonnet 4.6 | 89.6 |
| 9 | Grok 4.1 Fast | 89.6 |
| 10 | DeepSeek V3.2 | 88.1 |
| 11 | GPT-5.4 Nano | 87.3 |
Interactive version: theaggregate.ai/benchmark?slug=sake-five-shot-software-architecture-and-quantum-computing · How It Works · Data refreshed daily, snapshot 2026-09-19.