Gemini 3.8 LivevsGrok Voice 3.0

Grok Voice 3.0 is more reliable, by 40.2 points. Gemini 3.8 Live passed 34 of 82 scenarios on all three runs, and Grok Voice 3.0 passed 67. A scenario only counts when every run of it passed.

For a voice agent, Grok Voice 3.0, since it gets more scenarios right on every run. Grok Voice 3.0 is ahead of Gemini 3.8 Live on reliability, success rate, agent responses, data accuracy, calls where the agent went quiet and response time. The two tie on interruption.

Grok Voice 3.0, by 16.9 points. Gemini 3.8 Live saved everything exactly right on 70.5% of the calls that save data, and Grok Voice 3.0 on 87.3%. On Booking calls, 78.4% against 98.2%. On Medicare calls, 50.0% against 59.1%. A call counts only when nothing is missing, wrong, extra or saved twice.

No. Grok Voice 3.0 starts replying 0.57s sooner at the median. Gemini 3.8 Live takes 1.91s at the median and 2.23s at p90; Grok Voice 3.0 takes 1.34s and 1.62s. Both are measured on the call audio, from the caller finishing to the agent starting to reply.

They score the same. Gemini 3.8 Live scores 4.98 and Grok Voice 3.0 4.98 for how well the agent handled being talked over.

Grok Voice 3.0, by 2.0 points. The agent went quiet for 10 seconds or more on 2.0% of Gemini 3.8 Live calls and 0.0% of Grok Voice 3.0 calls.

There is no verified per-minute rate for Gemini 3.8 Live, so cost cannot be compared yet.

Cekura ran both as the whole agent in the same open-source Pipecat pipeline, with the same prompt, tools and connection. A simulated caller worked through all 82 scenarios on live calls, three times each, 3,198 calls across the benchmark. A call passes only when the agent said the right things and saved the right data.