Released: 2026-05-06
Key:
Score Threshold:
google/gemini-3-flash-preview
Score: 3561
Distance: 679km
Lat: 39.011663
Lng: 34.150537
The landscape features arid, rolling farmland with reddish-brown tilled soil, which is highly characteristic of the Central Anatolia Region of Turkey. The presence of slender poplar trees in the distance is a common sight in rural Turkish landscapes. The overall vegetation, lighting, and horizon suggest a location near the province of Kırőehir or Aksaray.
meta-llama/llama-4-maverick
meta
Score: 1359
Distance: 2606km
Lat: 41.216667
Lng: -4.583333
The image depicts a rural landscape with a plowed field, sparse vegetation, and trees in the distance. The presence of a Google watermark suggests it is a Street View image. The dry, brown field and the type of vegetation are indicative of a Mediterranean climate. The landscape and vegetation are consistent with regions in Spain, such as the province of Segovia, where the coordinates 41.216667, -4.583333 are located.
anthropic/claude-haiku-4.5
anthropic
Score: 2523
Distance: 1368km
Lat: 50.2
Lng: 14.4
This image shows a flat, semi-arid agricultural landscape with scattered vegetation, poplar trees in the distance, and dry grassland. The terrain and vegetation patterns are characteristic of Central European steppe-like regions, particularly the Czech Republic or neighboring areas in the continental climate zone. The presence of poplar trees (commonly planted windbreaks), the flat topography, sandy/loamy soil visible in the foreground, and the overall landscape composition suggest this is likely in central Bohemia or Moravia region of the Czech Republic, where such open agricultural areas with similar vegetation and tree types are common. The Google watermark and landscape characteristics point to this region around the 50°N, 14°E coordinates.
anthropic/claude-sonnet-4.6
anthropic
Score: 3867
Distance: 514km
Lat: 39.8
Lng: 32.5
The image shows a dry, flat agricultural landscape with plowed fields, sparse dry vegetation, and poplar trees in the background. The arid, semi-steppe appearance with the dry golden-brown grass and the style of vegetation (poplars, scrubby bushes) is characteristic of central Anatolia, Turkey. The flat terrain and dry summer conditions suggest this is somewhere in the Ankara or Konya region of Turkey.
openai/gpt-5-image-mini
openai
Score: 1394
Distance: 2554km
Lat: 39.412
Lng: -3.421
Flat, dry cereal fields with poplar windbreaks and sparse scrub are characteristic of the La Mancha plateau in central Spain. The low horizon, stubbled brown fields after harvest and isolated tree clumps match landscapes around Alcázar de San Juan / Ciudad Real region south of Madrid (central Spain).
x-ai/grok-4-fast
x-ai
Score: 1363
Distance: 2600km
Lat: 39.5
Lng: -4
The image depicts a rural, semi-arid landscape with dry, plowed fields covered in stubble, patches of dry grass, and clusters of tall, slender trees resembling poplars used as windbreaks. The clear blue sky, minimal vegetation, and distant hazy hills suggest a Mediterranean climate during a dry season, such as late summer or autumn. This scenery is characteristic of the La Mancha region in central Spain, known for its vast agricultural plains and similar terrain visible in Google Street View imagery.
moonshotai/kimi-k2.5
moonshot
Score: 1401
Distance: 2545km
Lat: 40
Lng: -3.5
The image shows a dry, plowed agricultural field with reddish-brown soil characteristic of the Iberian Peninsula. The vegetation includes scattered trees (likely poplars) and dry Mediterranean scrub, typical of central Spain’s Meseta region (Castilla-La Mancha or Castilla y León). The flat to gently rolling terrain, clear skies, and farming patterns match rural areas between Madrid and Toledo. The landscape is inconsistent with Northern Europe (too dry), the Americas (wrong vegetation types), or other Mediterranean regions due to the specific combination of soil color, field patterns, and tree species.
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.