OpenAI's GPT-Realtime-2.1-mini brings reasoning and tool use to cheaper voice agents

AI/OpenAI/
OpenAI's GPT-Realtime-2.1-mini brings reasoning and tool use to cheaper voice agents
Illustration

OpenAI released gpt-realtime-2.1 and a cheaper gpt-realtime-2.1-mini on July 6 for low-latency voice agents. The mini adds realtime reasoning and tool use at budget pricing, and can keep talking while running a tool call to avoid dead-air pauses. Across the Realtime voice models, OpenAI says improved caching cut p95 latency by at least 25%. Reasoning effort is configurable per turn.

3 sources

  1. 01OpenAI Releases GPT-Realtime-2.1 and GPT-Realtime-2.1-mini (MarkTechPost)marktechpost.com
  2. 02OpenAI Realtime API Cuts Voice Agent Latency 25%, Adds Reasoning Mini Model (TechTimes)techtimes.com
  3. 03OpenAI Releases GPT-Realtime-2.1 Voice Models With Lower Latency (Lets Data Science)letsdatascience.com