GPT-Live 1 + GPT-6 Sol (low thinking) is ahead on reliability, success rate, agent responses, data accuracy, response time and cost, and Deepslate Opal on stalled calls and interruption.
Last updated Methodology by Dileep Chagam
Pick Deepslate Opal for fewer stalled calls (6.9 points lower) and handling interruptions (0.07 higher). Pick GPT-Live 1 + GPT-6 Sol (low thinking) for reliability across repeated calls (18.3 points higher), calls that pass every check (13.8 points higher), getting the agent's replies right (4.1 points higher), saving the caller's data correctly (16.5 points higher), fast replies (0.25s sooner) and lower cost ($0.057 cheaper).
Scenarios that passed on all three runs.
GPT-Live 1 + Sol low, 18.3 points higher
Calls that passed every check.
GPT-Live 1 + Sol low, 13.8 points higher
Calls where the agent's replies reached the expected outcome.
GPT-Live 1 + Sol low, 4.1 points higher
Calls where everything the agent saved matched exactly.
GPT-Live 1 + Sol low, 16.5 points higher
Calls where the agent went silent for 10 seconds or more.
Deepslate Opal, 6.9 points lower
How well it handled being talked over, 0 to 5.
Deepslate Opal, 0.07 higher
Median time to start replying after the caller finishes.
GPT-Live 1 + Sol low, 0.25s sooner
One minute of call at the vendor's published rates.
GPT-Live 1 + Sol low, $0.057 cheaper
The small figure is the place among ranked models. The cascade is not ranked.
GPT-Live 1 + Sol low 2.02s, Deepslate Opal 2.27s at the median.
GPT-Live 1 + GPT-6 Sol (low thinking) is more reliable, by 18.3 points. Deepslate Opal passed 47 of 82 scenarios on all three runs, and GPT-Live 1 + GPT-6 Sol (low thinking) passed 62. A scenario only counts when every run of it passed.
For a phone agent, GPT-Live 1 + GPT-6 Sol (low thinking), since it gets more scenarios right on every run. GPT-Live 1 + GPT-6 Sol (low thinking) is ahead on reliability, success rate, agent responses, data accuracy, response time and cost, and Deepslate Opal on stalled calls and interruption.
GPT-Live 1 + GPT-6 Sol (low thinking), by 16.5 points. Deepslate Opal saved everything exactly right on 72.6% of the calls that save data, and GPT-Live 1 + GPT-6 Sol (low thinking) on 89.0%. On Appointments, 85.4% against 100.0%. On Medicare, 39.4% against 60.6%. A call counts only when nothing is missing, wrong, extra or saved twice.
No. GPT-Live 1 + GPT-6 Sol (low thinking) starts replying 0.25s sooner at the median. Deepslate Opal takes 2.27s at the median and 2.83s at p90; GPT-Live 1 + GPT-6 Sol (low thinking) takes 2.02s and 2.42s. Both are measured on the call audio, from the caller finishing to the agent starting to reply.
Deepslate Opal, by 0.07 on a 0 to 5 scale. Deepslate Opal scores 4.96 and GPT-Live 1 + GPT-6 Sol (low thinking) 4.89 for how well the agent handled being talked over.
Deepslate Opal, by 6.9 points. The agent went silent for 10 seconds or more on 5.3% of Deepslate Opal calls and 12.2% of GPT-Live 1 + GPT-6 Sol (low thinking) calls.
GPT-Live 1 + GPT-6 Sol (low thinking) is $0.057 cheaper a minute. Deepslate Opal costs $0.112 a minute of call and GPT-Live 1 + GPT-6 Sol (low thinking) $0.055, at each vendor's published rates.
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.
