Retell uses OpenAI models to handle more than 100 million calls a month
Retell uses OpenAI models for response generation, coding and interface testing across more than 100 million monthly voice-agent calls for over 4,000 businesses.
AT A GLANCE
Key points
- Retell reports more than 100 million calls processed each month.
- OpenAI models generate responses for 80% of calls.
- GPT‑6 Astra scored 89.7% in Retell’s vCX evaluation.
What is changing?
According to an OpenAI customer story, Retell processes more than 100 million calls a month on its platform for building enterprise voice AI agents and serves over 4,000 businesses. The company says OpenAI models generate responses for 80% of those calls.
Retell combines conversation, tool calls and human escalation within agents used for support, appointment and sales workflows. The company reports that its agents resolve more than 80% of customer issues and that the platform has processed over one trillion tokens; these are Retell’s own operational metrics.
GPT‑6 Astra scored 89.7% and ranked first in Retell’s vCX instruction-following evaluation. The team also uses Codex for research, development and debugging, and is testing GPT‑6 Astra computer use against front-end changes. These vendor-run tests should be treated as product-selection signals rather than an independent benchmark.
Explore the original sources
Source date: 27 October 2026.
Published by AI Radar: 28 Oct 2026, 02:08 (Istanbul, UTC+3).
FOLLOW THE TOPIC
Topics and tags
Share your discovery.
Send this story to a friend.
Reader comments
What do you think about this development? Share your experience or question. Comments are published after editorial approval.
Your name/nickname and comment become public if approved. Read: Privacy notice.
Loading comments…