Atlas

Benchmarks

← All benchmarks

HardGeoBench Avg Distance

Multimodal · 2025-05-08

HardGeoBench average geodesic distance in miles between a model's predicted photo location and the true location. Lower scores indicate more accurate geolocation; official leaderboard rows use pass@1 raw model calls without web search or tools.

Top models (lower is better)

ModelScore
Claude Opus 5.5655
Claude Sonnet 5.5683
Gemini 3.1 Pro Preview733
Gemini 4 Argon741
Gemini 3.6 Flash775
Claude Fable 5800
Gemini 3 Flash Preview842
Gemini 3.8 Flash843
Claude Fable 5.1845
Muse Spark 1.2871
Gemini 3.5 Flash886
Muse Spark 1.1889
Gemini 3.7 Flash895
GPT-6 Astra908
Muse Spark Open948
Gemini 3 Pro Preview990
Muse Spark 1.31022
Sonnet 4.61043
o31081
Gemini 2.5 Pro1119
Gemini 2.0 Flash1128
GPT-51158
Gemini 2.5 Pro Experimental 03-251172
GPT-5.51196
Gemini 2.0 Flash-Lite1216
Claude Opus 51218
GPT-6.1 Sol1239
Opus 4.81245
Opus 4.71279
GPT-5.11303
GLM-5.3 Flash1309
Claude Haiku 5.51319
o3-pro1340
Gemini 2.5 Flash Preview (09-2025)1343
Muse Wonder1369
GPT-5.6 Sol1381
Gemini 3.1 Flash-Lite Preview1435
Gemini 2.5 Flash1444
GPT-4.11494
GPT-6 Sol1497
Loading Atlas data…