GPT-Realtime-2.1 Official Specifications, Pricing & Benchmarks
GPT-Realtime-2.1 (gpt-realtime-2.1) by OpenAI, released 2026-07-06, featuring 128,000 context window, 32,000 max output tokens, and Text+Audio+Image input.
📋 GPT-Realtime-2.1 Official API & System Card Specifications
Verified against official OpenAI developer documentation & pricing
1. Identity & Capabilities
API Spec- API Model ID
- gpt-realtime-2.1
- Release Date
- 2026-07-06
- Weight License
- Proprietary (Undisclosed)
- Native Reasoning
- ✅ Supported
- Function Calling & JSON Schema
- 部分支持
- Batch API 异步批处理
- ✅ Supported (50% off)
2. Context, Output & Modalities
Limits- Max Context Window
- 128,000 tokens (128K)
- Max Output Tokens
- 32,000 tokens
- Input Modalities
- Text + Audio + Image
- Output Modalities
- Text + Audio
- Direct China Access
- 🌐 Global / OpenRouter
- Deployment Mode
- Cloud API only (Proprietary)
3. Official Pricing & Cache
USD / 1M- Input / 1M Tokens
- $ 4
- Cached Input / 1M
- $ 0.4 -90%
- Output / 1M Tokens
- $ 24
🎛️ Official Sampling & Reasoning Parameters
Recommended generation settings & API parameter constraints
💡 Custom temperature disabled in reasoning mode; adjust reasoning effort/budget instead.
Closest Self-Hostable Open-Weight Alternatives to GPT-Realtime-2.1
Open-weight models with full HuggingFace weights for private GPU / Mac deployment
Frequently Asked Questions
Is GPT-Realtime-2.1 open-source or proprietary?
GPT-Realtime-2.1 (gpt-realtime-2.1) is a proprietary commercial model by OpenAI. Weights and internal layer dimensions are undisclosed; it is accessed via cloud API.
What are the context limit and API pricing of GPT-Realtime-2.1?
GPT-Realtime-2.1 supports 128K input context and 32,000 max output tokens, priced at $4/1M input ($0.4/1M cached) and $24/1M output.