TravelEval - Scene-Time Ratio: leaderboard
Metric: Scene-Time Ratio (STR): effective visit time over total allocated time per attraction, including queuing (%); Direct prompting (one pass, no tools), plans simulated in the TravelEval sandbox of 10 Chinese cities (real rail, flight, hotel and attraction data, queuing-time model, road-network distances) over 1,150 queries; higher is better. Source: arxiv.org. Saturation forecast: Around August 2028. 7 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 2.0 Flash | 74.76 |
| 2 | Qwen 3 8B | 73.13 |
| 3 | GPT-4o Mini (2024-07-18) | 72.94 |
| 4 | DeepSeek V3.1 (Non-reasoning) | 70.23 |
| 5 | GPT-5 Chat | 69.59 |
| 6 | GPT-4o (2024-11-20) | 68.68 |
| 7 | Claude Sonnet 4.5 | 68.2 |
Interactive version: theaggregate.ai/benchmark?slug=traveleval-scene-time-ratio · How It Works · Data refreshed daily, snapshot 2026-09-29.