Cekura Bench

Cekura Bench publishes verifiable benchmarks for voice AI models. Its speech-to-speech benchmark evaluates 9 realtime voice models—including GPT Realtime 2.1, Gemini Live, Grok, and Phonic—as complete phone agents on live calls, covering 82 scenarios with three runs each. Models are ranked on reliability, data accuracy, stalled calls, response time, and cost, and every call transcript is public. The platform also covers voice agent and speech-to-text benchmarks, with text-to-speech benchmarks planned.