Released: 2026-05-12
Key:
Score Threshold:
meta-llama/llama-4-maverick
meta
Score: 3088
Distance: 964km
Lat: 22.722222
Lng: 88.363333
The image depicts a rural road with lush greenery on both sides, typical of certain regions in India. The presence of ’Google’ watermark in the bottom corners indicates it’s a Google Street View image. The road’s condition and surroundings suggest a less developed area. Given the vegetation and road characteristics, it is likely located in Eastern India. The specific coordinates (22.722222, 88.363333) point to a location in West Bengal, India, which matches the observed scenery.
google/gemini-3-flash-preview
Score: 3500
Distance: 713km
Lat: 26.549221
Lng: 88.086324
The vegetation, including tropical trees and flat agricultural fields, suggests a South Asian landscape. The road quality, the presence of a cyclist in the distance, and the specific lighting/camera quality are consistent with rural West Bengal or the Terai region in eastern Nepal and India near the border. The 2018-2020 copyright watermark often correlates with coverage in parts of the Indian subcontinent such as the Siliguri or Jhapa border areas.
anthropic/claude-sonnet-4.6
anthropic
Score: 3445
Distance: 745km
Lat: 22.5
Lng: 84.5
This image shows a narrow rural road through flat agricultural land with sparse trees and dry vegetation, consistent with the dry season in central/eastern India. The landscape, vegetation type, road construction, and hazy sky are characteristic of states like Jharkhand, Odisha, or Chhattisgarh in India. The red laterite soil visible on the roadside and the overall terrain suggest this is likely in the Jharkhand/Odisha region of India.
anthropic/claude-haiku-4.5
anthropic
Score: 3757
Distance: 572km
Lat: 30.8
Lng: 75.9
This image shows a rural road in what appears to be northern India, likely in Punjab or Haryana based on several clues: the flat agricultural landscape with scattered trees, the style of fencing, the reddish soil visible on the road, the tropical vegetation, and the general appearance of rural Indian infrastructure. The landscape is very flat with agricultural fields on both sides, typical of the Indo-Gangetic Plain. The road appears to be a country road connecting rural areas, common in Punjab region. The weather appears clear and warm, consistent with northern India’s climate. Based on these characteristics, this appears to be somewhere in the Punjab region of India, around the coordinates of rural areas near Ludhiana or Ferozepur district.
x-ai/grok-4-fast
x-ai
Score: 3881
Distance: 507km
Lat: 25.5941
Lng: 85.1376
The image depicts a narrow, paved rural road winding through flat agricultural fields, likely paddy fields, surrounded by sparse vegetation and trees typical of the Indian countryside. The hazy sky and tropical dry-season appearance suggest a location in the Gangetic plain of northern India, such as near Patna in Bihar, where such Street View imagery is common.
openai/gpt-5-image-mini
openai
Score: 4537
Distance: 194km
Lat: 26.775
Lng: 80.889
Rural, flat agricultural landscape in dry-season condition with narrow paved road, scrubby trees and small fields, simple single-wire utility poles and no road markings or curbs — features typical of northern Indian countryside (Uttar Pradesh/Haryana region). Vegetation and sun angle match subtropical plains rather than SE Asian rainforest or African savannah, so I place it in rural Uttar Pradesh around 26.8°N, 80.9°E.
moonshotai/kimi-k2.5
moonshot
Score: 4427
Distance: 243km
Lat: 26.4499
Lng: 80.3333
The image depicts a paved rural road through flat agricultural terrain with tropical/subtropical vegetation. The dense greenery on the left resembles bamboo thickets common in North India, while the scattered trees and agricultural fields on the right are characteristic of the Indo-Gangetic plain. The unpaved shoulders, road construction style, and vegetation patterns match rural Uttar Pradesh. The flat topography and presence of utility poles along the road further suggest this region. The lighting and haze indicate a warm climate consistent with the Gangetic plains of India.
Each model receives the same Google Street View image. Models are tasked with guessing the location of the image using only visual clues from the street view. Their answers are evaluated without feedback or retries.