Google has been continuously refining its AI chatbot, Gemini, since its launch. The latest upgrade, the Gemini 1.5 Pro-002 model, is now available to developers and paid subscribers through Gemini Advanced. This release follows the recent rollout of Gemini Live, which enables Android users to interact with the AI via voice.
The Gemini 1.5 Pro-002 model is designed to tackle math, coding, and complex queries with unmatched efficiency. This upgraded version outperforms its predecessor by facilitating deeper discussions and handling more intricate math equations.
According to Google, the new model demonstrates a 7% improvement in performance on the MMLU-Pro benchmark test, which assesses reasoning abilities.
The performance gains extend to the Math and HiddenMath benchmark tests, where the model showed a remarkable 20% increase in capability. Additionally, in Python Visual Comprehension and Code Generation tests, the model achieved improvements ranging from 2% to 7%. These enhancements promise to deliver more accurate answers to user queries.
Previously, Gemini received an upgrade to 1.5 Flash in July, along with significant latency improvements in August. Google also introduced custom chatbots, known as Gems, during the same month.
[embedpost slug="/comstech-and-huawei-collaborate-on-cybersecurity-workshop/"]
