Model Metadata
- type
- realtime
- name
- GPT-Realtime-2.1
- description
- Realtime speech-to-speech model with configurable reasoning, tool use, and robust voice-agent behavior
- family
- gpt
- open weights
- No
- knowledge
- 2024-09-30
- release date
- 2026-07-06
- modalities
- input: text, audio/pcm, audio/pcmu, audio/pcma, image/jpeg, image/png, image/gif, image/webp; output: text, audio/pcm, audio/pcmu, audio/pcma
- capabilities
- tools: supported: Yes; reasoning: supported: Yes; temperature: No; top_p: No; top_k: No
- limit
- context: 128,000; input: 96,000; output: 32,000
Provider Variants 2
OpenAI gpt-realtime-2.1
Context 128,000 · Output 32,000
- type
- realtime
- name
- GPT-Realtime-2.1
- description
- Realtime speech-to-speech model with configurable reasoning, tool use, and robust voice-agent behavior
- family
- gpt
- open weights
- No
- knowledge
- 2024-09-30
- release date
- 2026-07-06
- modalities
- input: text, audio/pcm, audio/pcmu, audio/pcma, image/jpeg, image/png, image/gif, image/webp; output: text, audio/pcm, audio/pcmu, audio/pcma
- capabilities
- tools: supported: Yes; reasoning: supported: Yes; effort: minimal, low, medium, high, xhigh; structured_output: No; temperature: No; top_p: No; top_k: No; system_updates: No; effort_updates: No
- limit
- context: 128,000; input: 96,000; output: 32,000
- id
- gpt-realtime-2.1
- canonical id
- openai/gpt-realtime-2.1
- last updated
- 2026-07-06
- api
- realtime: base_url: https://api.openai.com/v1; sdk: aisdk: @ai-sdk/openai; capabilities: tools: updates: No
vercel openai/gpt-realtime-2.1
Context 128,000 · Output 32,000
- type
- realtime
- name
- gpt-realtime-2.1
- description
- Realtime speech-to-speech model with configurable reasoning, tool use, and robust voice-agent behavior
- family
- gpt
- open weights
- No
- knowledge
- 2024-09-30
- release date
- 2026-07-06
- modalities
- input: text, audio/pcm, audio/pcmu, audio/pcma; output: text, audio/pcm, audio/pcmu, audio/pcma
- capabilities
- tools: supported: Yes; reasoning: supported: Yes; effort: minimal, low, medium, high, xhigh; cache: implicit: Yes; temperature: Yes; top_p: No; top_k: No; system_updates: No; effort_updates: No
- limit
- context: 128,000; input: 96,000; output: 32,000
- id
- openai/gpt-realtime-2.1
- canonical id
- openai/gpt-realtime-2.1
- last updated
- 2026-07-06
- api
- realtime: base_url: https://ai-gateway.vercel.sh/v4/ai; sdk: aisdk: @ai-sdk/gateway; capabilities: tools: updates: No