Retell Powers 100M Monthly Calls With OpenAI Models
The voice AI platform Retell now handles more than 100 million calls per month, utilizing advanced OpenAI models to maintain high call quality and drive impressive enterprise efficiency.
Radwan AbdulhadiEditor and technology writer
Rapid Scaling and Enterprise Adoption
Retell AI operates a voice AI platform tailored for contact center operations, enabling businesses to construct agents that can answer questions, take action, and seamlessly connect callers with human representatives when necessary. More details can be found via [Retell AI (opens in a new window)](https://www.retellai.com/).
The platform currently serves more than 4,000 businesses. Monthly call volume on the platform climbed from 30 million calls in early 2026 to surpass 100 million calls by September. To handle complex enterprise workflows, Retell relies on OpenAI models to preserve high call quality and reliability. At present, OpenAI models generate responses for 80 percent of calls on Retell, where customers routinely achieve resolution rates above 80 percent. Furthermore, Retell has crossed the threshold of 1 trillion tokens consumed across OpenAI models.
Performance Benchmarks and Model Evaluation
Because enterprise voice agents must adhere strictly to company policies and maintain designated roles despite shifting conversation paths, Retell designed its vCX benchmark. This benchmark evaluates dozens of models using demanding multi-turn conversations derived from actual production scenarios, testing instruction following, information grounding, tool calling, and natural conversational language.
In Retell’s October 13, 2026 benchmark evaluation, GPT-6 Astra ranked first in instruction following with a score of 89.7 percent. Additional evaluations conducted prior to the launch of GPT-6 Astra demonstrated that GPT-5.5 achieved the highest rankings among tested models for both non-hallucination rate at 88.0 percent and tool-call correctness at 88.3 percent when reasoning was enabled, outperforming Gemini-3.1-Pro and Claude-4.6 on both parameters.
Real-World Enterprise Deployments
Contact center teams utilize Retell to establish business context, goals, and rules for their agents, allowing developers to further refine workflows and integrations. During active calls, Retell manages audio models, tracks conversations, and controls when agents speak, listen, or handle interruptions. Models such as GPT-6.1 Sol, GPT-6 Astra, and GPT-5.5 generate responses and request tool calls using customer instructions and contextual data.
For electronics company Anker, Retell agents manage Tier-1 post-sales support across the United States and the United Kingdom, covering tasks like order tracking, warranty checks, and troubleshooting. Anker reported an 80.4 percent case resolution rate along with a Net Promoter Score of 63. Meanwhile, Pine Park Health deployed Retell agents for appointment reminders and proactive scheduling, successfully recovering medical-assistant time equivalent to 2.1 full-time roles.
Infrastructure, Latency, and Engineering Efficiency
Daytime traffic on Retell's platform can surge to more than 50 times its overnight low. To maintain responsiveness during peak demand, Retell uses dedicated servers shared across its customers in coordination with OpenAI. In the October 13, 2026 benchmark, GPT-6 Astra and GPT-6 Sol both recorded median first-sentence latencies of 883 milliseconds, while GPT-5.6 Luna achieved a top ranking at 497 milliseconds.
Engineers across the organization also integrate OpenAI Codex throughout the development lifecycle, using it to research unfamiliar code, construct features, and debug issues. This streamlined development workflow has allowed the company to scale its revenue effectively, reaching a $1.6 million annual revenue run rate per employee.
Future Outlook and Platform Expansion
Retell continues to expand its platform capabilities, with future plans focused on enhancing quality assurance and simulation tools, incorporating support for additional languages, and developing knowledge base auto-learning features. Startups looking to scale similar applications can explore opportunities through [Start building (opens in a new window)](https://openai.com/startups).
As OpenAI models continue to evolve, Retell expects its enterprise customers to deliver sophisticated service experiences to millions of callers on their own terms. Additional information regarding OpenAI developments and enterprise scaling can be reviewed at [OpenAI News](https://openai.com/index/retell-ai/).
Sources
- OpenAI NewsRetell uses OpenAI models to help power 100M monthly calls
Continue chronologically
Related entity coverage
Related stories
OpenAI Launches GPT-6.1 Terra With 3x Faster AI Agents
OpenAI has officially launched GPT-6.1 Terra, a new model focused on computer use, visual reasoning, coding, and everyday work with 50% lower pricing.
Sullivan & Cromwell Develops Agreement Analyzer with OpenAI
Sullivan & Cromwell has collaborated with OpenAI to develop Agreement Analyzer, an AI tool designed to identify issues and review draft agreements.
Proaction boosts sales 60% and saves 75+ hours with Codex
Discover how Proaction uses OpenAI's Codex and GPT models to build, operate, and sell modern fleet management software faster.
Airbnb expands access to GPT-6 Astra for engineering teams
Airbnb is widening access to OpenAI frontier models, including GPT-6 Astra, to help engineering teams solve bugs, design systems, and ship features faster.
