CekuraBench

GPT-Live 1 + GPT-6 Sol

Every modelCompare

75.6%

Reliability

3rd of 12

2.02s

Median reply

7th of 12

$0.055

Cost / min

2nd of 9

Reliability vs. response time

Higher and further left is better

Best quadrant
Reliability, pass³ (%)
Median response time (s)

For the wider field, see our pick of the best voice AI APIs for real-time audio.

GPT-Live 1 + GPT-6 Sol against every other model

Scenarios that passed on all 3 runs, pass³. Higher is better.

#ModelReliability
  1. 01Grok Voice 3.081.7%
  2. 02GPT Realtime 2.179.3%
  3. 03GPT-Live 1 + GPT-6 Sol75.6%
  4. 04Azure Realtime70.7%
  5. 05Phonic v168.3%
  6. 06Grok Voice Think Fast 2.067.1%
  7. 07Gemini 3.1 Flash Live Preview62.2%
  8. 08Deepslate Opal57.3%
  9. 09Gemini 3.8 Live Extended Thinking52.4%
  10. 10Amazon Nova 2 Sonic50.0%
  11. 11GPT Realtime 2.1 Mini50.0%
  12. 12Gemini 3.8 Live41.5%
  • refCascade82.9%

Where GPT-Live 1 + GPT-6 Sol struggled

GPT-Live 1 + GPT-6 Sol is at or above the median on every kind of call.

GPT-Live 1 + GPT-6 Sol is at or above the median on every kind of call.

Mistakes in saved data
  • Wrong value22 · 52%
  • Step skipped18 · 43%
  • Field left out2 · 5%

Fields it got wrong most: destination, route_status, state

Response time at P50, P90 and P95

GPT-Live 1 + GPT-6 Sol starts replying in 2.02s on a typical turn and 2.59s on the slowest turn in twenty, 7th fastest of 12 at the median.

ModelP50P90P95
  • Cascade
    2.17s2.51s2.62s
Seconds from the caller finishing to the agent replying. The bar is the typical turn, P50.

GPT-Live 1 + GPT-6 Sol head to head

Every comparisonAll 12 models

How much audio GPT-Live 1 + GPT-6 Sol was tested on

246 live calls: 82 scenarios, each run three times. The median call ran 1:51.

Live calls
246
Scenarios
82
Booking calls
177
Medicare calls
69
Replies per call
9.6
Median call
1:51

82 scenarios, the same for every model. See every test case

GPT-Live 1 + GPT-6 Sol FAQ

  • GPT-Live 1 + GPT-6 Sol passed 62 of 82 scenarios on all three runs, 75.6%, 3rd of 12 realtime models. Across all 246 calls, 83.7% passed every check.

  • GPT-Live 1 + GPT-6 Sol starts replying 2.02s after the caller finishes at the median, 2.42s at P90 and 2.59s at P95. That is 7th fastest of 12 at the median, measured on the call audio.

  • GPT-Live 1 + GPT-6 Sol is at or above the field median on every kind of call. When it saved data wrong, the most common mistake was wrong value, 52% of them.

  • GPT-Live 1 + GPT-6 Sol saved everything exactly right on 89.0% of the calls that save data, 2nd of 12. On Booking calls, 100.0%. On Medicare calls, 60.6%.

  • $0.055 a minute of call at OpenAI's published rates, 2nd cheapest of 9 priced models. Includes the delegated backend model.

  • Cekura ran GPT-Live 1 + GPT-6 Sol as the whole agent in the same open-source Pipecat pipeline as every other model, with the same prompt, tools and connection. A simulated caller worked through 82 scenarios on live calls, three times each: 246 calls for GPT-Live 1 + GPT-6 Sol. A call passes only when the agent said the right things and saved the right data.