Google stands by Gemini 4 performance as some claim it ‘struggles’ in real-world use

Google responded to a Bloomberg report that some employees found Gemini 4's real-world performance inconsistent with its benchmark scores, particularly in coding tasks.

Google stated it would be inaccurate to say Gemini 4 is underperforming in coding, and a DeepMind executive reaffirmed the model is at the frontier.

The report notes a split among employees, with some confident in the model's frontier status and others concerned it lags behind competitors.