Azure Realtime

70.7%

Reliability

5th of 14

1.66s

Median reply

5th of 14

$0.080

Cost / min

6th of 10

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.

Azure Realtime against every other model

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

#ModelReliability
  1. 01Phonic v190.2%
  2. 02Grok Voice 3.081.7%
  3. 03GPT Realtime 2.179.3%
  4. 04GPT-Live 1 + GPT-6 Sol75.6%
  5. 05Azure Realtime70.7%
  6. 06Prism68.3%
  7. 07Grok Voice Think Fast 2.067.1%
  8. 08Gemini 3.1 Flash Live Preview62.2%
  9. 09Deepslate Opal58.5%
  10. 10Higgs Realtime56.1%
  11. 11Gemini 3.8 Live Extended Thinking52.4%
  12. 12Amazon Nova 2 Sonic50.0%
  13. 13GPT Realtime 2.1 Mini50.0%
  14. 14Gemini 3.8 Live41.5%
  • refCascade82.9%

Where Azure Realtime struggled

Azure Realtime falls furthest behind on background noise: 67% passed, against a median of 81%.

Background noise: 67% passed, median 81%Background noise67%median 81%Several requests in one call: 86% passed, median 95%Several requests in one call86%median 95%Plain booking & lookup: 96% passed, median 96%Plain booking & lookup96%median 96%Silence, cough & packet loss: 92% passed, median 88%Silence, cough & packet loss92%median 88%Safety & scope: 89% passed, median 83%Safety & scope89%median 83%Tool or system errors: 100% passed, median 92%Tool or system errors100%median 92%Caller corrects or changes mind: 92% passed, median 79%Caller corrects or changes mind92%median 79%Interruptions & barge-in: 100% passed, median 87%Interruptions & barge-in100%median 87%Medicare compliance & routing: 65% passed, median 48%Medicare compliance & routing65%median 48%Accent & slow speech: 100% passed, median 83%Accent & slow speech100%median 83%
  • Background noise67% median 81%
  • Several requests in one call86% median 95%
  • Plain booking & lookup96% median 96%
  • Silence, cough & packet loss92% median 88%
  • Safety & scope89% median 83%
  • Tool or system errors100% median 92%
  • Caller corrects or changes mind92% median 79%
  • Interruptions & barge-in100% median 87%
  • Medicare compliance & routing65% median 48%
  • Accent & slow speech100% median 83%
Azure RealtimeMedian of all modelsInside the dashed line means fewer calls passed than most models.
Mistakes in saved data
  • Field left out45 · 55%
  • Step skipped14 · 17%
  • Wrong value12 · 15%

Fields it got wrong most: route_id, lead_id, route_status

How fast Azure Realtime replies

Half of Azure Realtime's replies start within 1.66s of the caller finishing, and 19 in 20 within 2.42s. That is 5th fastest of 14.

Half of replies

P50

within 1.66s

9 in 10 replies

P90

within 2.22s

19 in 20 replies

P95

within 2.42s

Prism

Fastest model · P50

within 1.34s
The gap is the silence after the caller stops talking, before Azure Realtime starts its reply.CallerAzure Realtime

How much audio Azure Realtime was tested on

246 live calls: 82 scenarios, each run three times. The median call ran 1:59.

Live calls
246
Scenarios
82
Booking calls
177
Medicare calls
69
Replies per call
7.0
Median call
1:59

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

Azure Realtime FAQ

  • Azure Realtime passed 58 of 82 scenarios on all three runs, 70.7%, 5th of 14 realtime models. Across all 246 calls, 84.1% passed every check.

  • Azure Realtime starts replying 1.66s after the caller finishes at the median, 2.22s at P90 and 2.42s at P95. That is 5th fastest of 14 at the median, measured on the call audio.

  • Azure Realtime falls furthest behind the field on background noise: 67% of calls passed, against a median of 81%. When it saved data wrong, the most common mistake was field left out, 55% of them.

  • Azure Realtime saved everything exactly right on 86.1% of the calls that save data, 5th of 14. On Booking calls, 94.2%. On Medicare calls, 65.2%.

  • $0.080 a minute of call at Microsoft's published rates, 6th cheapest of 10 priced models. Each call's metered tokens at Voice Live's Pro rates, read from Azure's price list on 7 Oct 2026. Excludes caller transcription, which Azure bills separately as speech input.

  • Cekura ran Azure Realtime 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 Azure Realtime. A call passes only when the agent said the right things and saved the right data.

Test your voice agent the same way.

Cekura runs these simulated calls on your agent and scores every one.