OrderProbe - Chain-of-Thought: leaderboard
Metric: Exact Recovery Rate (%; chain-of-thought prompt). Source: arxiv.org. Saturation forecast: Around November 2027. 12 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | DeepSeek V3.2 | 62.24 |
| 2 | Gemini 2.5 Flash | 41.56 |
| 3 | GPT-4o | 37.69 |
| 4 | Qwen 3 14B | 31.53 |
| 5 | DeepSeek R1 | 27.73 |
| 6 | Llama 3 8B Instruct | 24.43 |
| 7 | Claude Haiku 4.5 | 21.65 |
| 8 | Llama 2 7B Chat | 18.75 |
| 9 | Gemma 3 4B (IT) | 18.03 |
| 10 | Qwen 3 8B | 12.59 |
Interactive version: theaggregate.ai/benchmark?slug=orderprobe-chain-of-thought · How It Works · Data refreshed daily, snapshot 2026-09-25.