In a surprising turn of events for Shakespeare enthusiasts, Google has rebranded Bard as Gemini, introducing a more advanced version called Gemini Advanced. Drawing parallels to ChatGPT’s base model and premium ChatGPT Plus service, Gemini and Gemini Advanced offer enhanced capabilities at a higher price point.
During my extensive testing of generative AIs over the past year, both Google and OpenAI have charged $20/month for access to their advanced offerings. In my coding challenges, ChatGPT consistently outperformed Google’s Bard, which struggled to meet basic requirements on multiple occasions. Even Meta’s touted Code Llama AI fell short in comparison.
Despite suggestions from readers to vary the challenges presented to AI models, I have maintained a consistent set of tests to directly compare performance levels. While ChatGPT excelled in a variety of programming languages and tasks, Bard’s shortcomings prevented further evaluation until it could handle fundamental tests.
Now, with the rebranding of Bard to Gemini and the introduction of Gemini Advanced, it’s time to reassess Google’s AI capabilities with a fresh set of challenges. The initial test involved writing a simple WordPress plugin, a task that Bard failed twice previously. Gemini Advanced attempted to create a front-end feature, showcasing its proficiency in both PHP and JavaScript. However, functionality issues persisted, prompting a second attempt with mixed results.
In another test to rewrite a string function, Gemini Advanced struggled with basic concepts of dollars and cents processing, highlighting significant limitations in its programming capabilities. Similarly, the AI’s inability to identify and resolve a bug further underscored its shortcomings compared to ChatGPT’s problem-solving proficiency.
As the AI landscape continues to evolve, the performance disparities between different models become increasingly evident. While ChatGPT has consistently demonstrated strong performance across various tasks, Google’s Gemini Advanced faces significant challenges in meeting basic coding requirements. The quest for AI excellence in programming remains ongoing, with each test revealing new insights into the capabilities of these advanced systems.