04/30/2025
New research by Mobi.AI data scientist Jiahao Liu finds that Mobi’s AI outperforms Gemini Pro and GPT-4 at travel itinerary generation by over 30x.
A recent benchmark study from Fudan University, The Ohio State University, The Pennsylvania State University, and Meta AI found that LLMs (large language models) like Google’s Gemini Pro and OpenAI’s GPT-4 perform poorly in travel planning, with success rates of only 1.1% and 0.6%, respectively.
As part of Mobi.AI’s latest research to support our Intent Driven Search product, Mobi has built a prototype system to explore how AI can more intelligently and reliably plan multi-step travel itineraries. Mobi’s AI outperforms existing models significantly, achieving a 32.8% success rate—over 30 times higher than other LLMs.
Learn more about the why behind this significant difference in our article: https://www.mobi.ai/resources/the-future-of-travel-planning-how-ai-is-revolutionizing-your-itinerary
Data scientist Jiahao Liu explains how even the most advanced AI models today struggle with itinerary planning, but how new research from Mobi is revolutionizin...