AI Model Comparison: GPT-5's Geolocation Capabilities Lag Behind Other Artificial Intelligence Models
In a recent AI geolocation test conducted by Bellingcat, involving 25 of their own holiday photos sourced from every continent, Google AI Mode stood out as the most capable tool overall.
At the time, OpenAI's ChatGPT o4-mini-high emerged as the clear winner, outperforming most other models, including Google Lens. However, in the latest tests, Google AI Mode outperformed even the Gemini 2.5 Pro Deep Research model in geolocation accuracy.
The test involved 24 models, including GPT-5 Pro and Thinking, Grok 4, and older models such as GPT, Claude, Gemini, and Grok. Most of these models, including GPT-5, were found to be less accurate than Google AI Mode, with many incorrectly identifying a beach in France.
Interestingly, Google AI Mode was the only model to correctly identify the location as Noordwijk, Netherlands, in Test 25. This was a significant improvement over the older models, which accurately identified the country but failed to locate the town in Test 25.
The results of the geolocation tests were ranked on a scale from 0 to 10, with 10 indicating an accurate and specific identification. Google's AI Mode scored high marks, demonstrating its accuracy and reliability in this area.
It's worth noting that OpenAI removed the option to select older models such as o4-mini-high after the release of GPT-5. However, after negative feedback, OpenAI reinstated GPT-4o as the default model for paid subscribers.
Despite delivering faster answers, GPT-5 appeared to sacrifice accuracy, as reported by other users. This was particularly noticeable in the geolocation tests, where GPT-5 often failed to provide accurate results.
The test also revealed that the majority of models, at some point, returned a hallucination in the geolocation tests. This highlights the need for continued development and improvement in AI models to ensure they provide accurate and reliable results.
Google describes AI Mode as its "most powerful AI search, with more advanced reasoning and multimodality." Currently, AI Mode is available in 180 countries worldwide but remains unavailable in the European Union and Germany, likely due to complex EU regulations such as data protection laws.
Bellingcat will continue to test new models as they emerge in the field of AI geolocation. The latest tests have shown that while there is room for improvement, Google AI Mode is currently the most reliable and accurate tool in this area.
Read also:
- Understanding Hemorrhagic Gastroenteritis: Key Facts
- Trump's Policies: Tariffs, AI, Surveillance, and Possible Martial Law
- Expanded Community Health Involvement by CK Birla Hospitals, Jaipur, Maintained Through Consistent Outreach Programs Across Rajasthan
- Abdominal Fat Accumulation: Causes and Strategies for Reduction
