ChronoBench - Temporal Recognition: leaderboard
Recognize which land-cover transitions a class or a located object underwent between two given dates or over the whole sequence, and whether a class expanded or contracted; the temporal recognition level of ChronoBench (2026). Accuracy in percent, higher is better.
Metric: Accuracy (%; level average of bi-temporal class-level change and area-change recognition, long-temporal class-level change recognition and bi-temporal object-level change recognition; 17,689 validated multiple-choice questions over remote-sensing image time series, box- and geo-coordinate-grounded variants; open models greedy with 128 new tokens, API models at high reasoning effort). Source: arxiv.org. Saturation forecast: Around February 2027. 11 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.4 (High) | 67.57 | #123 (GPT-5.4) |
| 2 | Gemini 3 Flash | 61.38 | #164 |
| 3 | Qwen 3 VL 32B Instruct | 57.88 | #585 |
| 4 | Qwen 3 VL 8B Instruct | 47.12 | #981 |
| 5 | Qwen 3 VL 4B Instruct | 42.07 | #1279 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=chronobench-temporal-recognition · How It Works · Data refreshed daily, snapshot 2026-11-13.