Cekura S2S - Pass^3: leaderboard
Cekura Bench realtime speech-to-speech leaderboard: each voice model runs as the agent on 82 scripted phone scenarios (59 patients booking, moving or cancelling clinic visits; 23 Medicare callers whose eligibility is checked and who are routed), three times each, inside the same open-source Pipecat agent with the same prompt and tools, over live WebRTC audio including noisy and glitchy calls. Labels carry the model setting Cekura states. Score: pass^3 (%), the share of the 82 scenarios the model passed in all three runs, a consistency measure stricter than the per-call success rate. Higher is better.
Metric: Pass^3 (%): scenarios passed in all three runs, of 82. Source: benchmarks.cekura.ai. Saturation forecast: Rough model projection: around 2026. 13 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Phonic 1 (Intelligence high) | 90.2 | |
| 2 | Cekura Cascade (flux-general-en → gpt-4.1 → eleven_flash_v2_5) | 82.9 | |
| 3 | Grok Voice 3.0 (Reasoning high) | 81.7 | |
| 4 | GPT Realtime 2.1 (Reasoning high) | 79.3 | |
| 5 | GPT-Live 1 + GPT-6 Sol (Sol on low thinking) | 75.6 | |
| 6 | Azure Realtime | 70.7 | |
| 7 | Grok Voice Think Fast 2.0 (Reasoning high) | 67.1 | |
| 8 | Gemini 3.1 Flash Live Preview (Thinking level high) | 62.2 | |
| 9 | Deepslate Opal | 57.3 | |
| 10 | Gemini 3.8 Live Extended Thinking (Thinking level high) | 52.4 |
Interactive version: theaggregate.ai/benchmark?slug=cekura-s2s-pass-pow-3 · How It Works · Data refreshed daily, snapshot 2026-11-05.