AI vs Puzzles

Back to Home
Puzzle#248

Geoguessr

Released: 2026-07-10

Google Street View Image
Country Flag

Thailand

Lat: 18.97569515043934

Lng: 99.21726541369904

See on Google Street View


6
Models Tested
1621.7km
Avg. Distance
3296.8
Avg. Score
5
LLM Providers

Key:

Attempt
Solution

Score Threshold:

>4400
>2000
<2000

google/gemini-3-flash-preview

google

Score: 4617

Distance: 159km

Lat: 18.243542

Lng: 97.917424


The image displays a tropical landscape with lush, forested hills in the background and simple wooden or corrugated metal structures typical of rural Southeast Asia. The dry, scrubby terrain in the foreground and the specific style of the vegetation and utility poles are highly characteristic of northern Thailand or eastern Myanmar. The geographic layout matches areas in the Mae Hong Son province of Thailand near the Salween River Valley.


meta-llama/llama-4-maverick

meta

Score: 4433

Distance: 241km

Lat: 16.883333

Lng: 98.633333


The image appears to be a rural or countryside scene with a mix of trees and cleared land, and a building in the distance. The architecture and vegetation suggest a location in Southeast Asia. The presence of a dirt field and the style of the building are consistent with regions in northern Thailand or Myanmar. Given the landscape and climate visible in the image, a plausible location is near the Thailand-Myanmar border. The coordinates 16.883333, 98.633333 are in Mae Hong Son Province, Thailand, which fits the observed environment.


anthropic/claude-haiku-4.5

anthropic

Score: 1810

Distance: 2032km

Lat: 21.1458

Lng: 79.8882


This image shows a rural settlement in hilly terrain with dry deciduous forest characteristics typical of central India. The landscape features red/laterite soil, sparse vegetation in dry season, scattered houses with basic construction, and rolling hills with forest cover. The vegetation pattern, soil color, terrain, and settlement style are consistent with the Vidarbha region of Maharashtra state or similar areas in central India. The dry season appearance and forest type suggest this is likely in the Yavatmal or Wardha district area of Maharashtra, in the heart of the Deccan plateau region.


anthropic/claude-sonnet-4.6

anthropic

Score: 3831

Distance: 533km

Lat: 14.5

Lng: 101


The image shows a rural landscape with dry scrubby vegetation, scattered trees, low wooden structures, and forested hills in the background. The dry season appearance, vegetation type, and architectural style of the simple rural buildings are consistent with rural Thailand. The terrain with hills in the background and the overall appearance suggests central or western Thailand, possibly in Kanchanaburi or nearby province area.


openai/gpt-5-image-mini

openai

Score: 173

Distance: 6732km

Lat: -19.6

Lng: 146.8


Rural, subtropical landscape with dry grass, open eucalypt-like trees and simple tin-roofed buildings on stilts — vegetation and house style are typical of inland coastal Queensland, Australia. The low serrated ridgeline in the background resembles the Great Dividing Range foothills near the Townsville/Ingham/Cardwell region. Power poles and fencing match Australian rural infrastructure, so I’ve placed the guess in northern Queensland (around 19.6°S, 146.8°E).


MoonshotAI

moonshotai/kimi-k2.5

moonshot

Score: 4917

Distance: 33km

Lat: 18.7936

Lng: 98.9658


The image depicts a rural landscape characteristic of northern Thailand, specifically the Chiang Mai region. Key indicators include: (1) tropical/subtropical vegetation with deciduous trees showing seasonal dryness, (2) rolling forested hills in the background typical of the mountainous terrain in this region, (3) simple rural structures with corrugated metal roofs common in Thai countryside, (4) dry season grass conditions suggesting a monsoon climate with distinct wet/dry seasons, and (5) the specific type of scrubland and scattered trees matches the agricultural landscape found in the valleys and foothills around Chiang Mai province. The power lines and general infrastructure level also align with rural Thai development patterns.


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