Model Details
Released
May 7, 2026
Context Length
128,000
Max Output
32,000
Where is gpt-realtime-2 a perfect fit?
GPT-Realtime-2 is OpenAI's realtime voice model with GPT-5-class reasoning, designed for more complex speech-to-speech interactions, stronger instruction following, and reliable tool use.
- Sophisticated speech-to-speech assistants requiring low latency
- Customer-support, sales, booking, and call-center voice agents
- Voice agents that need reasoning before responding or taking actions
- Tool-driven workflows involving function calls and external systems
- Long-running conversations requiring better context handling
- Applications needing natural spoken interaction rather than speech-to-text → LLM → TTS pipelines
- Multimodal agents combining audio, text, and image inputs
- Voice-to-action applications that interpret requests and execute appropriate tools
- Sophisticated speech-to-speech assistants requiring low latency
- Customer-support, sales, booking, and call-center voice agents
- Voice agents that need reasoning before responding or taking actions
- Tool-driven workflows involving function calls and external systems
- Long-running conversations requiring better context handling
- Applications needing natural spoken interaction rather than speech-to-text → LLM → TTS pipelines
- Multimodal agents combining audio, text, and image inputs
- Voice-to-action applications that interpret requests and execute appropriate tools
Quick Model Estimate
Your GPT-5 Cost Estimate
💰 Total Cost
USD 3.00
for 1000 input + 1000 output tokens
📥 Input (1000 × $4.000000)
USD 1.5000
📤 Output (1000 × $24.000000)
USD 1.5000
Cost Breakdown
📥 Input 50%
📤 Output 50%
Prices updated daily from official provider data.
Pricing
|
Provider
↕
|
Modality
↕
|
Service tier
↕
|
Input Price
(per 1M tokens) ↕ |
Output Price
(per 1M tokens) ↕ |
Context Size
↕
|
View
|
|---|---|---|---|---|---|---|
 OpenAI | Text | Standard | $4.0000 | $24.0000 | 128,000 tokens | → |
| OpenAI | Audio | Standard | $32.0000 | $64.0000 | 128,000 tokens | → |
| OpenAI | Image | Standard | $5.0000 | N/A | 128,000 tokens | → |
FAQs about gpt-realtime-2
How much does gpt-realtime-2 cost per 1M tokens?
gpt-realtime-2 costs between $4.00 and $32.00 per million input tokens, and between $24.00 and $64.00 per million output tokens.
What does a typical workload cost with gpt-realtime-2?
1,000 requests of 2,000 input and 500 output tokens each — 2,000,000 input and 500,000 output tokens in total — costs $20.00 with gpt-realtime-2 at its lowest rates. Use the calculator on this page for your own volumes.
Where can I use gpt-realtime-2?
gpt-realtime-2 is available through OpenAI, from $4.00 per million input tokens.
What is gpt-realtime-2's context window?
gpt-realtime-2 has a context window of 128,000 tokens, and returns up to 32,000 tokens in a single response.