GPT Realtime 2.1 is ahead on reliability, success rate, agent responses and data accuracy, and Gemini 3.8 Live on stalled calls and response time. The two tie on interruption.
Last updated Methodology by Dileep Chagam
Pick Gemini 3.8 Live for fewer stalled calls (1.2 points lower) and fast replies (0.03s sooner). Pick GPT Realtime 2.1 for reliability across repeated calls (37.8 points higher), calls that pass every check (19.9 points higher), getting the agent's replies right (4.9 points higher) and saving the caller's data correctly (21.5 points higher). They are level on handling interruptions.
Scenarios that passed on all three runs.
GPT Realtime 2.1, 37.8 points higher
Calls that passed every check.
GPT Realtime 2.1, 19.9 points higher
Calls where the agent's replies reached the expected outcome.
GPT Realtime 2.1, 4.9 points higher
Calls where everything the agent saved matched exactly.
GPT Realtime 2.1, 21.5 points higher
Calls where the agent went silent for 10 seconds or more.
Gemini 3.8 Live, 1.2 points lower
How well it handled being talked over, 0 to 5.
Level
Median time to start replying after the caller finishes.
Gemini 3.8 Live, 0.03s sooner
One minute of call at the vendor's published rates.
Not comparable
The small figure is the place among ranked models. The cascade is not ranked.
Gemini 3.8 Live 1.91s, GPT Realtime 2.1 1.94s at the median.
Both providers, every model: Google vs OpenAI
GPT Realtime 2.1 is more reliable, by 37.8 points. Gemini 3.8 Live passed 34 of 82 scenarios on all three runs, and GPT Realtime 2.1 passed 65. A scenario only counts when every run of it passed.
For a phone agent, GPT Realtime 2.1, since it gets more scenarios right on every run. GPT Realtime 2.1 is ahead on reliability, success rate, agent responses and data accuracy, and Gemini 3.8 Live on stalled calls and response time. The two tie on interruption.
GPT Realtime 2.1, by 21.5 points. Gemini 3.8 Live saved everything exactly right on 70.5% of the calls that save data, and GPT Realtime 2.1 on 92.0%. On Appointments, 78.4% against 95.9%. On Medicare, 50.0% against 81.8%. A call counts only when nothing is missing, wrong, extra or saved twice.
Yes. Gemini 3.8 Live starts replying 0.03s sooner at the median. Gemini 3.8 Live takes 1.91s at the median and 2.23s at p90; GPT Realtime 2.1 takes 1.94s and 2.46s. 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 GPT Realtime 2.1 4.98 for how well the agent handled being talked over.
Gemini 3.8 Live, by 1.2 points. The agent went silent for 10 seconds or more on 2.0% of Gemini 3.8 Live calls and 3.3% of GPT Realtime 2.1 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, 2,952 calls across the benchmark. A call passes only when the agent said the right things and saved the right data.
