TIDE Trajectory - Sudoku - Success: leaderboard
Metric: Success Rate (%). Source: arxiv.org. 17 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 2.5 Pro | 99 |
| 2 | GPT-OSS-120B | 97 |
| 3 | Gemini 2.5 Flash (Non-reasoning) | 97 |
| 4 | DeepSeek R1 | 95 |
| 5 | DeepSeek V3.2 (Non-reasoning) | 93 |
| 6 | GLM-4 32B (0414) | 55 |
| 7 | Phi-4 (Reasoning) | 51 |
| 8 | Llama 3.3 70B Instruct | 40 |
| 9 | Phi-4 | 33 |
| 10 | Llama 3.1 8B Instruct | 4 |
| 11 | GLM-4 9B Chat | 4 |
Interactive version: theaggregate.ai/benchmark?slug=tide-trajectory-sudoku-success · How It Works · Data refreshed daily, snapshot 2026-09-19.