A comparison of Google's Gemini and Anthropic's Claude for drafting email replies found that Claude was better at matching tone and requirements. The writer used Gemini 3.5 Flash through a Pro account and the free Claude Sonnet 4.6 model with Effort set to High. For each email, the writer outlined the draft's objective and gave two standard instructions: asking the AI to suspend immediate responses and refrain from follow-up questions, and commanding it to imitate the writer's tone from the email thread. In a test using a prompt about a shipment delay and missing tracking number, Gemini asked two follow-up questions, including a legitimate detail about the targeted completion date. The writer noted that Claude's speech-to-text engine is leagues ahead of Gemini, especially if you have an accent.