Side-by-side comparison

Pick up to 4 models. Differences are what matter for picking.

Models
Spec
OpenAI
GPT Realtime
Verified 2026-05-08
DescriptionSpeech-to-speech model for low-latency voice agents. Native audio understanding and generation over the Realtime API.
Released2025-08-28
Knowledge cutoff2024-09
Context window32K
Max output
Input $/1M$4.00
Output $/1M$16.00
Speed tierUltra Fast
Reasoning3/5
Coding3/5
Multilingual5/5
Modalities in
Text
Audio In
Modalities out
Text
Audio Out
Tool useYes
JSON modeYes
Open weightsNo
Best use cases
Voice
Chat
Agents
Best personas
Consumer
Developer Tools
Startup
Weak at
  • · long-context reasoning
  • · complex coding
Compliance
HIPAA-eligibleSOC 2GDPRISO 27001FedRAMP
Data policyAPI inputs/outputs not used to train OpenAI models by default.
Sources
Tip: hop over to the Recommender to describe a problem and have AI suggest the best fit from this dataset.