TIDE Trajectory - AlfWorld - Success: leaderboard
Metric: Success Rate (%). Source: arxiv.org. 17 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | DeepSeek R1 | 83.6 |
| 2 | Gemini 2.5 Pro | 80.7 |
| 3 | DeepSeek V3.2 (Non-reasoning) | 80.7 |
| 4 | Gemini 2.5 Flash (Non-reasoning) | 78.6 |
| 5 | Llama 3.3 70B Instruct | 59.3 |
| 6 | GPT-OSS-120B | 50.7 |
| 7 | GLM-4 32B (0414) | 37.9 |
| 8 | GLM-4 9B Chat | 25.7 |
| 9 | Phi-4 | 25 |
| 10 | Llama 3.1 8B Instruct | 16.4 |
| 11 | Phi-4 (Reasoning) | 15.7 |
Interactive version: theaggregate.ai/benchmark?slug=tide-trajectory-alfworld-success · How It Works · Data refreshed daily, snapshot 2026-09-19.