gpt-realtime-2.1-mini
by OpenAI · Released 2026
Distilled, lower-cost gpt-realtime-2.1 for faster, cheaper realtime voice interactions. Same 2.1 improvements at a fraction of the audio cost.
gpt-realtime-2.1-mini
Powered by OpenAI · Realtime multimodal
Context Window
128K
Parameters
Not disclosed
Max Output
N/A
Category
LLM Chat
Overview
`gpt-realtime-2.1-mini` is the distilled, lower-cost variant of gpt-realtime-2.1 — same recognition, silence/interruption, and reasoning improvements at a fraction of the audio cost. 128K text context. Voice-agent only (WebSocket).
Audio in/out rates are much lower than the full 2.1 ($10/$20 per 1M audio); text output is $2.40/1M. On CallMissed it bills against active call minutes — roughly $0.117/min — making it the cheapest premium realtime option.
Use it for high-volume voice agents where cost matters but you still want the 2.1 recognition and interruption quality.
Pricing
| Metric | Price |
|---|---|
| Input /1M tokens | ₹60.0000 |
| Output /1M tokens | ₹240.0000 |
1 credit = ₹1 = $0.01 USD. Prices shown from provider; CallMissed passes through with ~35% markup.
Key Highlights
- Low-cost 2.1 realtime
- Same recognition + interruptions
- $0.117/min
Technical Details
- Model id: gpt-realtime-2.1-mini
- Voice-agent WebSocket only
- ~$0.117 per active call minute
Strengths
- Cheapest premium realtime
- 2.1 recognition quality
- 128K context
Limitations
- Not available on chat completions
- Lower ceiling than full 2.1
Use Cases
API Example
# Create a voice session with llm_model=gpt-realtime-2.1-mini via POST /v1/voice/sessions
Endpoint: WebSocket /v1/voice/sessions · Model ID: gpt-realtime-2.1-mini
Try gpt-realtime-2.1-mini now
Get 1000 free API credits on signup. No credit card required.