AI vs Puzzles

Back to Home
Puzzle#246

Geoguessr

Released: 2026-07-08

Google Street View Image
Country Flag

France

Lat: 44.22260610353251

Lng: 2.534233473416614

See on Google Street View


6
Models Tested
1344.5km
Avg. Distance
3775.8
Avg. Score
5
LLM Providers

Key:

Attempt
Solution

Score Threshold:

>4400
>2000
<2000

google/gemini-3-flash-preview

google

Score: 4940

Distance: 24km

Lat: 44.202302

Lng: 2.234327


The landscape features rolling hills with a mixture of agricultural fields and small groves of deciduous trees, typical of the Aveyron or Tarn regions in Southern France. The presence of a concrete utility pole with specific wiring attachments and a small white box is characteristic of French rural infrastructure. The dry grass suggests a Mediterranean-influenced climate, and the distant stone farm buildings match the architectural style of the Occitanie region.


anthropic/claude-haiku-4.5

anthropic

Score: 3337

Distance: 809km

Lat: 43.2

Lng: 12.5


This image shows rolling agricultural hills with mixed farming and vegetation patterns typical of central Italy, specifically the Tuscany or Umbria region. The landscape features gentle rolling terrain with scattered trees, cultivated fields, and patches of vegetation characteristic of the Val d’Orcia or similar areas in southern Tuscany. The utility pole and the general terrain composition suggest a rural area in central Italy. The Google Earth watermark and the pastoral landscape are consistent with the Tuscany region around latitude 43.2 and longitude 12.5.


anthropic/claude-sonnet-4.6

anthropic

Score: 4660

Distance: 141km

Lat: 44

Lng: 0.8


The image shows rolling green hills with deciduous trees, scattered farm buildings in the distance, and a utility pole with what appears to be a small electrical box. The landscape is characteristic of rural southwestern France - the Lot-et-Garonne or Gers department area. The gentle hills, mixed farmland, and the style of the terrain are typical of the Gascony region. The vegetation and sky suggest a temperate climate in summer.


openai/gpt-5-image-mini

openai

Score: 4785

Distance: 88km

Lat: 44.5

Lng: 1.5


Rolling, gently undulating farmland with mixed pasture and small fields, hedgerows and scattered stone or stucco farmhouses in the distance. Deciduous trees and dry summer grass suggest a temperate, slightly continental/southern inland European climate. The concrete utility pole and small meter box are consistent with rural France, particularly the Lot/Dordogne/Haute-Vienne style countryside in southwest/central France. I therefore place the image in rural southwest France (approx. Occitanie/Nouvelle-Aquitaine border region).


meta-llama/llama-4-maverick

meta

Score: 158

Distance: 6913km

Lat: 38.9453

Lng: -84.5441


The landscape appears to be a rural area in the United States with rolling hills and a mix of grass and trees. The presence of a utility pole and wires suggests a typical North American infrastructure. The date stamp ’2015’ on the image and the vegetation suggest a temperate climate. Based on the layout of the fields and trees, it is likely in the Midwest or Eastern United States. A possible match is in Kentucky, with coordinates near 38.9453, -84.5441.


MoonshotAI

moonshotai/kimi-k2.5

moonshot

Score: 4775

Distance: 92km

Lat: 44.8

Lng: 1.7


The image shows a rural landscape with rolling hills, mixed forest and pasture land, and a wooden utility pole with electrical equipment. The terrain and vegetation are characteristic of the French countryside, particularly the regions of Occitanie or Nouvelle-Aquitaine in southwestern France. The style of the utility pole with the attached junction box and the general agricultural landscape patterns are consistent with rural French infrastructure. The dry grass suggests late summer timing. Without distinctive landmarks, this is an approximate guess based on the topography and rural character matching the limestone causse regions or agricultural areas of central-southern France.


Methodology Note

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.

Leaderboard