Model comparison
Compare GPT-Realtime-2.1 and Gemini 3.1 Flash Live Preview using the same provider-sourced voice & realtime rubric. No mystery score and no invented benchmark ranking.
Facts checked September 4, 2026
Quick take
OpenAI's realtime speech-to-speech reasoning model for tool-using voice agents that also need text and image context.
There is no published universal latency number, and audio-token costs are hard to compare directly with minute- or character-priced competitors.
Google's preview audio-to-audio model for low-latency dialogue with multimodal awareness, thinking, search grounding, and function calling.
The model is preview, and paid versus unpaid projects have different data-use terms. Confirm both before capturing sensitive conversations.
Compare the published facts
Values use each provider's own published units and limits. A blank means the provider did not publish a directly comparable value in the sources reviewed.
| Voice & realtime | GPT-Realtime-2.1 | Gemini 3.1 Flash Live Preview |
|---|---|---|
| Price basisToken-, character-, or minute-based price for the listed access route. | Audio $32 in · $64 out / 1M tokens | Audio $0.005/min in · $0.018/min out |
| Interaction modeSpeech-to-speech, audio-to-audio, or streamed text-to-speech behavior. | Speech-to-speech with text/image context | Realtime audio-to-audio, multimodal input |
| Published latencyProvider-published measurement with its stated exclusions; not a Cody test. | Not published | Not published |
| LanguagesProvider-documented language coverage or a clear unpublished marker. | Multilingual; exact list not published here | Multilingual; verify current Live API support |
| Tools & controlNative function calling, reasoning, or speech-control features. | Function calling and configurable reasoning | Function calling and search grounding |
| Context or input limitDocumented token or character limit where it is meaningful. | 128K input · 32K max output | 131,072 input · 65,536 output |
How to choose
Start with the job you need to complete, then validate cost, access, and policy details on your exact provider route.
OpenAI says API content is not used for training by default. Default abuse-monitoring logs may be retained up to 30 days, with additional controls available to qualifying organizations.
Provider and API links
Google says paid Gemini API content is not used to improve its products; unpaid service data generally can be. Confirm billing status and current terms before sending sensitive conversations.
Provider and API links
Frequently asked questions
GPT-Realtime-2.1: OpenAI's realtime speech-to-speech reasoning model for tool-using voice agents that also need text and image context. Gemini 3.1 Flash Live Preview: Google's preview audio-to-audio model for low-latency dialogue with multimodal awareness, thinking, search grounding, and function calling.
Consider GPT-Realtime-2.1 when your priority is Tool-using customer or employee voice agents. Consider Gemini 3.1 Flash Live Preview when your priority is Multimodal voice assistants. Test both with your own data and provider route before committing.
No. This comparison aligns provider-published facts for the Voice & realtime category. It does not claim a universal winner or combine incompatible third-party benchmark scores.
