OpenAI Realtime API

Own OpenAI Realtime API? Claim this listing

OpenAI's own speech-to-speech model API for voice agents: WebRTC, WebSocket or native SIP telephony, billed purely on audio tokens with no platform fee.

Visit OpenAI Realtime API Compare

Verified 2026-08-23 · theagentsindex.com

Category: Voice agents
From: $32 / $64 audio / per 1M tokens
Pricing model: usage-based
Free tier: No
Deployment: Cloud
Interface: API
Open source: No
Public API: Yes
Model / LLM: OpenAI gpt-realtime-2.1 (flagshi…

Facts re-verified 2 days ago

Our verdict

The OpenAI Realtime API is the rawest option in this category: not an orchestration platform or a managed dashboard, but the actual speech-to-speech model some of this category's own platforms let you plug in as one provider choice among several (LiveKit Agents' plugin ecosystem lists it directly).

Going straight to OpenAI trades every convenience those platforms bundle (a dashboard, built-in phone numbers, call analytics, a knowledge base) for direct, usage-based token pricing with no markup and no monthly platform fee, backed by OpenAI's own SOC 2 Type 2 and ISO 27001-family compliance.

The tradeoffs are real: sessions cap at 60 minutes, there's no dashboard or built-in CRM, and a developer writes their own tool-calling, session-management and observability code.

Pick it over a platform like Vapi or Retell AI when the priority is owning the integration layer entirely and being comfortable single-vendor on OpenAI; pick a platform instead when the priority is a working agent without writing that plumbing.

What is OpenAI Realtime API?

The OpenAI Realtime API is OpenAI's own speech-to-speech model API for building real-time voice agents: one model call handles listening, reasoning, tool use and speaking, rather than chaining separate speech-to-text, LLM and text-to-speech services together.

OpenAI first previewed it at DevDay in October 2024. The original gpt-realtime model reached general availability on August 28, 2025, alongside native SIP telephony, remote MCP server support and image input.

A second-generation gpt-realtime-2 with configurable reasoning effort shipped May 7, 2026, with gpt-realtime-translate and gpt-realtime-whisper added on dedicated /v1/realtime/translations and /v1/realtime/transcription_sessions endpoints at the same time.

The current-generation gpt-realtime-2.1 and the smaller gpt-realtime-2.1-mini shipped July 6, 2026. The API connects over WebRTC for browser and mobile clients, WebSocket for server-side pipelines, or SIP for real phone calls, and bills purely on audio and text tokens rather than a platform subscription or a per-minute rate.

What does OpenAI Realtime API do?

Developers open a session against the Realtime API, authenticated with a short-lived ephemeral client secret rather than a long-lived API key, and choose gpt-realtime-2.1 (flagship, with configurable reasoning effort via reasoning.effort) or gpt-realtime-2.1-mini (faster, lower cost) as the model.

Turn detection is on by default using semantic_vad, which decides when a caller has stopped speaking to trigger a response and handles interruptions; it can be disabled entirely for manual, push-to-talk-style control instead. Mid-conversation, the model can call developer-defined functions (look up an order, check a calendar, pull account context) and speak the result back once the client returns it.

Ten built-in voices are available: alloy, ash, ballad, coral, echo, sage, shimmer, verse, marin, and cedar. OpenAI recommends marin or cedar for quality. The output voice cannot be changed once the model has produced its first audio response. gpt-realtime-2 and newer accept image input alongside audio. Every session is capped at 60 minutes, a hard platform limit other vendors abstract away.

There is no dashboard, no built-in phone number and no bundled telephony vendor: connecting over SIP for real inbound or outbound calls, wiring tool calls into a CRM or database, and building session-management, logging and observability are all code the developer writes directly against the API. Translation and transcription sessions use dedicated endpoints, billed by audio duration rather than tokens.

How OpenAI Realtime API works

Open a session against the Realtime API using a short-lived ephemeral client secret from POST /v1/realtime/client_secrets (no long-lived API key in the browser); WebRTC connection helper: POST /v1/realtime/calls
Choose a model. gpt-realtime-2.1 is the flagship reasoning-capable option. gpt-realtime-2.1-mini is the faster, lower-cost distilled version. Use a separate endpoint for translation (/v1/realtime/translations) or transcription (transcription sessions)
Connect over WebRTC (browser and mobile), WebSocket (server-side audio pipelines) or SIP (real phone calls). Set turn_detection: semantic_vad for automatic turn-taking with interruption handling, or disable it for manual control
The model calls developer-defined functions mid-conversation (look up an order, check a calendar) and speaks the result back once the client returns it. Configure a truncation.retention_ratio below 1 to control context size and cache behaviour
Sessions cap at 60 minutes; billing is purely usage-based. Voice-agent sessions accrue audio and text input plus output tokens. Translation and transcription sessions accrue audio duration instead

Key features

Native speech-to-speech model: One model call handles listening, reasoning, tool use and speaking. No separate STT, LLM and TTS chain to assemble and keep in sync. Configurable reasoning effort on gpt-realtime-2 and newer.
Three connection modes, including native SIP: WebRTC for browser and mobile clients, WebSocket for server pipelines, or SIP for real phone calls. No separate telephony vendor is required for the connection itself.
Server-side session controls: truncation.retention_ratio and token_limits.post_instructions cut input cost and shape what is sent to the model each turn. Prompt caching is automatic and works best when conversation history is static.
Mid-conversation function calling: The model invokes developer-defined functions live during a call and speaks the result back once the client returns it. Webhooks and server-side controls can drive a session from outside the client.
Configurable turn detection: semantic_vad is on by default: the model decides when the caller has stopped speaking and handles interruptions automatically. Can be disabled for manual, push-to-talk-style control.
Image input on Realtime 2 and newer: gpt-realtime-2, gpt-realtime-2.1 and their mini variants accept image input alongside audio, so an agent can reason over what is on screen or in a photo mid-call.
Separate translation and transcription endpoints: Dedicated /v1/realtime/translations and /v1/realtime/transcription_sessions endpoints, billed on audio duration rather than tokens, for streaming translation and transcription workflows.

What are OpenAI Realtime API's use cases?

Custom voice agent built without a platform: An engineering team writes its own session, tool-calling and telephony-handoff code directly against the model, paying only for tokens used with no platform markup on top.
The model layer beneath an orchestration framework: A team building on LiveKit Agents or a similar framework plugs OpenAI's Realtime API in as the speech-to-speech provider, keeping the framework's session and turn-detection plumbing while OpenAI's model does the listening, reasoning and speaking.
Inbound phone support over native SIP: A support line connects directly over SIP with no separate telephony vendor, with the model verifying the caller and pulling account context via function calls before escalating to a human when needed.

Who is OpenAI Realtime API for?

Engineering teams that want to go straight to OpenAI's own speech-to-speech model with no orchestration platform in between, and are comfortable owning the integration layer themselves
Teams already building on an orchestration framework (like LiveKit Agents) who want to plug OpenAI's model in as the speech-to-speech provider rather than assembling a separate STT/LLM/TTS pipeline
Buyers who want transparent, usage-based token pricing with no platform markup, and don't need a dashboard, call analytics or CRM integrations bundled in

Not forNon-technical teams who want a working voice agent without writing code, or anyone who wants CRM integrations, call analytics or a dashboard out of the box: a managed platform like Vapi, Retell AI, Bland AI or Synthflow AI fits that need better.

What does OpenAI Realtime API integrate with?

SIP trunking (native — real inbound/outbound phone calls with no separate telephony vendor required)
Model Context Protocol (MCP) — connects to remote MCP servers for external tool access
No native CRM, helpdesk or dashboard integrations — a developer wires session events, function calls and any downstream system into their own client/server code

Why use OpenAI Realtime API?

Direct access to OpenAI's own speech-to-speech model with no orchestration platform's markup, and the model keeps improving as OpenAI ships new Realtime versions. Four generations since first preview (gpt-realtime, gpt-realtime-1.5, gpt-realtime-2, gpt-realtime-2.1) in under two years.
Native SIP telephony, WebRTC and WebSocket support all ship directly in the API. Real inbound and outbound phone calls with no separate telephony vendor required, alongside browser and server clients.
Purely usage-based token pricing, published per model. No monthly platform subscription and no per-minute markup layered on top of the model cost.
Backed by OpenAI's own API-platform compliance stack (SOC 2 Type 2, ISO/IEC 27001:2022, ISO/IEC 27017, ISO/IEC 27018, ISO/IEC 27701) rather than a smaller vendor's still-building certification list.
Server-side session controls live in the protocol itself: truncation ratio, token limits and automatic prompt caching are first-class, configurable knobs on every session.

What are OpenAI Realtime API's pros and cons?

What's great

The only entry in this category priced purely in raw model tokens with no platform fee or per-minute markup. gpt-realtime-2.1 audio $32/$64 per 1M tokens in/out; gpt-realtime-2.1-mini $10/$20.
Native SIP telephony, WebRTC and WebSocket support all ship directly in the API. Real phone calls with no separate telephony vendor required for the connection itself.
OpenAI's API Platform (which the Realtime API is part of) carries SOC 2 Type 2, ISO/IEC 27001:2022, ISO/IEC 27017, ISO/IEC 27018 and ISO/IEC 27701 certification.
Session-level cost controls built into the protocol. Truncation ratio and token-limit settings plus automatic prompt caching let a developer cap input cost without external tooling.

Watch-outs

Sessions cap at 60 minutes. A hard platform limit other vendors in this category abstract away, so a long-running call needs a session-handoff strategy the developer builds themselves.
No dashboard, call analytics, CRM integrations or dedicated free tier. Every session, tool call and downstream integration is code a developer writes and maintains, unlike this category's managed platforms.
Single-vendor on OpenAI's own models only. Unlike Vapi, Retell AI or LiveKit Agents there is no bring-your-own-model option if a team wants a different LLM or voice provider.
Flagship reasoning model is real-time expensive at scale. gpt-realtime-2.1 audio output is $64 per 1M tokens, and reasoning.effort: high measurably increases latency and output token usage per turn.
60-minute maximum session duration.
No dashboard, call analytics or CRM / helpdesk integrations. A developer builds and maintains all of that.
No dedicated free tier for the Realtime API.
Single-vendor on OpenAI's own models only. No bring-your-own-model option for the speech-to-speech layer (the underlying LLM does run OpenAI models only).
32,000-token context window on the flagship model: a long-running call needs a truncation strategy to keep cost and latency stable.

OpenAI Realtime API pricing

Purely usage-based, billed on audio and text tokens (and on audio duration for the translation and transcription endpoints). No monthly platform fee and no dedicated free tier for the Realtime API itself. gpt-realtime-2.1: audio $32/$64 per 1M tokens in/out (cached audio input $0.40), text $4/$24 per 1M tokens in/out (cached text input $0.40). The smaller gpt-realtime-2.1-mini: audio $10/$20 per 1M tokens in/out (cached audio input $0.30), text $0.60/$2.40 per 1M tokens in/out (cached text input $0.06). OpenAI's Realtime conversations guide recommends keeping conversation history static to maximise prompt-cache hit rate; with caching in production, audio/text activity of roughly 600 input plus 1,200 output audio tokens per minute works out to an estimated $0.06-$0.11/min on the flagship model, or roughly a third of that on the mini model. Cost rises with conversation length because the conversation is appended to the prompt on each turn, with a 28,224-token headroom before the model's 32,000-token context forces truncation at a 4,096 max output.

gpt-realtime-2.1
$32 / $64 audio / per 1M tokens
- Audio input: $32 per 1M tokens ($0.40 cached input)
- Audio output: $64 per 1M tokens
- Text input: $4 per 1M tokens ($0.40 cached input)
- Text output: $24 per 1M tokens
- Image input: $5 per 1M tokens ($0.50 cached input)
- 128K-token context window, 32K max output tokens
- Configurable reasoning effort via `reasoning.effort`
- Estimated $0.06-$0.11/min in production once prompt caching is working (independently reported)
gpt-realtime-2.1-mini
$10 / $20 audio / per 1M tokens
- Audio input: $10 per 1M tokens ($0.30 cached input)
- Audio output: $20 per 1M tokens
- Text input: $0.60 per 1M tokens ($0.06 cached input)
- Text output: $2.40 per 1M tokens
- Image input: $0.80 per 1M tokens ($0.08 cached input)
- Distilled reasoning for faster, lower-cost realtime voice
- About a third of the flagship model's audio cost at the same conversation length
Translation endpoint (gpt-realtime-translate)
Custom, contact sales (per audio minute)
- Billed on audio duration through `/v1/realtime/translations`, not on Response tokens
- Live translation of streaming audio: WebRTC for browser ingest or WebSocket for server audio pipelines
- See Realtime pricing page for current translation rates
Transcription endpoint (gpt-realtime-whisper / gpt-live-transcribe)
Custom, contact sales (per audio minute)
- Billed on audio duration through `/v1/realtime/transcription_sessions`
- Streaming transcript deltas committed in real time
- See Realtime pricing page for current transcription rates

See current pricing on developers.openai.com ↗Compare OpenAI Realtime API alternatives →

Frequently asked questions

What is the OpenAI Realtime API?

OpenAI's own speech-to-speech model API for building real-time voice agents. One model call handles listening, reasoning, tool use and speaking. It connects over WebRTC, WebSocket or native SIP telephony. It is a model API, not a packaged voice-agent product: there is no dashboard, built-in phone number or CRM integration, and a developer writes the session, tool-calling and telephony code themselves.

How much does the OpenAI Realtime API cost?

Purely usage-based, billed on audio and text tokens (and on audio duration for the translation and transcription endpoints, which are billed separately). gpt-realtime-2.1: audio $32 input / $64 output per 1M tokens, text $4 / $24 per 1M tokens (cached audio $0.40, cached text $0.40). gpt-realtime-2.1-mini: audio $10 / $20 per 1M tokens, text $0.60 / $2.40 per 1M tokens (cached audio $0.30, cached text $0.06). At roughly 600 input and 1,200 output audio tokens per minute, with caching working, independently reported production usage lands around $0.06-$0.11/min on the flagship model and roughly a third of that on the mini. Cost depends heavily on conversation length.

Does the OpenAI Realtime API support phone calls?

No. The OpenAI Realtime API connects directly over SIP for real inbound and outbound phone calls, alongside WebRTC for browser and mobile clients and WebSocket for server-side pipelines. No separate telephony vendor like Twilio is required for the SIP path, though many production deployments still use one for number provisioning and call routing.

Is there a free tier for the OpenAI Realtime API?

No dedicated free tier or free minute allowance specific to the Realtime API. New OpenAI platform accounts have at various times received a small, one-time trial credit usable across all API products, but this is not Realtime-specific, is not guaranteed, and 2026 reports describe it as unreliable. Budget for paid usage from the first session.

Can I control token usage and prompt caching on a Realtime session?

Yes. Server-side session-management controls include a configurable truncation.retention_ratio and token_limits.post_instructions setting. The Realtime API also supports prompt caching automatically, and OpenAI's own guidance is to keep conversation history static to maximise cache hit rate and reduce input cost. Caching applies when subsequent turns' input tokens match tokens from an earlier turn, and it is best-effort rather than guaranteed.

The evidence

How we checked this

OpenAI's own model API, not a packaged voice-agent product: evaluate it as infrastructure, not a dashboard.

OpenAI Realtime API alternatives

LiveKit Agents
An open-source orchestration framework that can plug OpenAI's Realtime API in as one of several speech-to-speech providers, adding session/turn-detection plumbing and a self-hostable deployment path this raw model API doesn't provide on its own.
ElevenLabs Conversational AI
Starts from voice quality (10,000+ voices, instant cloning) and lets you bring any LLM including OpenAI's, rather than OpenAI's own single-vendor, usage-priced model with a fixed set of ten voices.
Compare the whole category
Every agent in Voice agents, side by side on the same fields.

Visit OpenAI Realtime APIOpens the vendor's own site Compare OpenAI Realtime APISide by side with the category See alternativesOther options worth weighing

Anything to add?

This page is researched and re-verified by our editors — but the people who actually use OpenAI Realtime API know things we cannot see from the outside.

Report an issueSomething wrong, outdated or a dead link? Tell us and we re-verify.

Write a reviewUsed OpenAI Realtime API? Your experience helps the next buyer decide.

Get a quote from OpenAI Realtime APITell us what you need and we make the introduction.

Get weekly Voice agents updates

New agents and fresh verdicts. One short email, every week. One-click unsubscribe.