14
AI Models Tested
30
Test Images
30
Countries Covered
1934-2024
Year Range
Benchmark Overview
Models are tasked with predicting both the year and geographic location where historical photographs were captured, using the official TimeGuessr scoring algorithm with a maximum of 10,000 points per image.
Geographic Localization
Models predict geographic coordinates (latitude, longitude) where historical images were captured
- • Distance-based scoring using Haversine formula
- • Visual cues: architectural styles, vegetation, license plates
- • Cultural and social indicators analysis
- • Maximum 5,000 points for perfect location accuracy
Temporal Estimation
Models predict the year when historical images were captured (range: 1800-2024)
- • Year-based scoring with absolute difference penalties
- • Temporal cues: technology, fashion, vehicles, image quality
- • Maximum 5,000 points for exact year prediction
- • Mean absolute error analysis for year predictions
Top Performing Models
Current leaderboard leaders based on TimeGuessr benchmark performance
#1
gemini-2.5-pro
Visiongemini-2.5-progoogle • Avg Score: 9587 pts
9587
Average Points (out of 10,000)
13.3%
Location Accuracy
1.8
Year MAE
Evaluated on 30 images using TimeGuessr scoring
#2
gpt-5
Visiongpt-5openai • Avg Score: 9364 pts
9364
Average Points (out of 10,000)
6.7%
Location Accuracy
2.6
Year MAE
Evaluated on 30 images using TimeGuessr scoring
#3gpt-4o
Visiongpt-4oopenai • Avg Score: 9241 pts
9241
Average Points (out of 10,000)
3.3%
Location Accuracy
2.4
Year MAE
Evaluated on 30 images using TimeGuessr scoring