Engadget Comparison Finds ChatGPT Voice More Expressive, Gemini Live Broader and Cheaper

Both chatbots followed a similar path into voice. When ChatGPT launched in 2022, users could only interact through a keyboard, but real-time voice chat soon became a breakthrough feature. ChatGPT Voice Mode was the first of its kind, and Google later introduced Gemini Live. Unlike older voice assistants such as Google Assistant, these chatbots can hold natural conversations on many subjects and search the internet for current information, according to Engadget.

On paper, the two services share many similarities. Both respond in real time, offer multiple voices, answer questions, look up information and switch between languages on the fly. The main technical difference is the underlying model family: ChatGPT uses GPT models, while Google's chatbot uses Gemini. Voice conversations also require fast responses, so both services use slightly less capable models that prioritize speed over accuracy, meaning responses are not of the same quality as those in each chatbot's text mode.

The comparison described a clear split in speaking style. ChatGPT Voice is much more expressive in its word selection and uses natural conversational markers, such as pretending to think with 'mhmm' and adding mid-sentence stutters. Gemini Live often sounds flatter in delivery because of differences in tone and inflection. Some users dislike ChatGPT's exaggerated intonations, Engadget noted, so they may prefer Google's approach.

Web search behavior was another difference. ChatGPT Voice is more likely to search the internet before answering. In one test, the author asked both chatbots about the recent Gemini 3.8 Live release. Gemini defaulted to its existing knowledge and said the release did not exist, while ChatGPT Voice paused for a couple of seconds, accessed the internet and ultimately gave the correct answer. In the author's experience, users who value factual accuracy often have to explicitly ask Gemini Live to research before it responds.

Free access is limited on both services. Users who have not paid for ChatGPT or Gemini quickly find that each chatbot offers only a few minutes of conversation before cutting access, because voice processing is significantly more computationally demanding than text. OpenAI and Google both sell paid plans that greatly increase the limit and grant access to more capable models. On ChatGPT's free tier, the author noticed that the GPT-Live-1 mini model lacks the reasoning skills of the larger GPT-Live-1 model. ChatGPT Go, the most affordable paid ChatGPT plan, costs $8 a month, while Google's AI Plus subscription costs $5 a month. The next step up for both is $20 a month. The author preferred Google's subscription options because even the lowest tier bundles 400GB of cloud storage and can be shared with up to five family members.

Both chatbots also support camera input alongside voice, allowing users to point a phone at their surroundings for added context. The author used Gemini Live while tuning up a bike, and the AI walked through the process and overlaid helpful markers on the live video feed so the user could quickly identify which screws needed tightening. ChatGPT's camera access during voice chat is locked behind its $20 monthly subscription tier. Google's equivalent feature is free, at least up to a certain limit. The entry-level ChatGPT Go subscription also lacks camera access during voice chat, requiring the $20 tier for that feature.

Google's breadth of hardware and software services gives it a natural advantage over OpenAI in voice access, according to the comparison. Gemini Live can be used on a dedicated smart speaker through Google Home Speaker, and Google has brought the feature to older Nest Audio and Hub devices. The author's Google Home Mini supports the feature, which the review called impressive for a device launched in 2017. The caveat is that an active Google Home Premium subscription at $10 a month is required.