Tasks and inputs
Eleven Realtime v1 Mini is listed for Audio, Audio-to-audio. It accepts audio, text and produces audio, text. Use the publisher’s input format and preprocessing for your chosen deployment.
SyncDev model directory · Profile updated Nov 17, 2026
Eleven Realtime v1 Mini is ElevenLabs' fast, low-cost speech-to-speech voice agent stack, with an OpenAI Realtime-compatible API and 90+ languages.
The specification score summarizes recorded features and limits. It is not a benchmark result.
Pricing conditions: About $0.08 per minute plus LLM usage. Indicative; confirm on your ElevenLabs plan.
Recorded modality support shows what this model can accept and produce. An undocumented modality is shown as unknown rather than assumed to be unsupported.
Accepts
Produces
Eleven Realtime v1 Mini is not one model in the usual sense. ElevenLabs describes it as a complete voice-agent configuration: speech recognition, turn-taking, a language model and text-to-speech, chosen and tuned together. You stream audio in and get spoken answers back. The underlying LLM is not disclosed.
Of the two Realtime configurations, this is the default, the fastest and the cheaper to run, tuned for natural conversation rather than heavy task execution.
Assistants, tutors and interactive characters: products where the feel of the conversation matters more than multi-step reasoning. Turn-taking and interruption handling are built in, voices come from the ElevenLabs library, and more than 90 languages are supported. Tool calling is available too, but if most turns involve chained lookups, its bigger sibling is the intended choice.
ElevenLabs designed the Realtime API to mirror OpenAI's Realtime API events and WebSocket transport. Existing client code written for GPT-Realtime-2.1 should need only a change of endpoint, credentials and model, though you should verify every event type your client depends on. It is also available inside ElevenAgents.
| Field | Value |
|---|---|
| Model ID | eleven_realtime_v1_mini |
| Type | Speech-to-speech agent configuration |
| Input / output | Audio and text / audio and text |
| Languages | 90+ |
| Tool calling | Yes |
| Reasoning before reply | No |
| Indicative price | $0.08 per minute plus LLM usage |
The per-minute rate covers the voice stack. LLM usage is billed on top, and since the LLM is not named, the only way to know your real cost per minute is to run real conversations and read the bill. Treat $0.08 as a floor, not a total.
The rollout started on 19 October 2026 and reached every workspace on 2 November. If you tried in late October and could not select it, that is why.
ElevenLabs claims 96% on a conversational-dynamics measure and 620 ms time to first audio for Mini. The latency number is end to end for the whole stack, not the speech model alone, so it is the one users will actually feel. Your network and region will move it.
Script twenty conversations that resemble your product, including interruptions, long pauses and a language switch. Have testers run them live, timing first audio after they stop speaking and noting every awkward overlap. Then pull the bill for those sessions and divide by minutes. If the agent stumbles on tasks that need several tool calls in a row, switch the model to v1 Max; ElevenLabs makes that a configuration change, not a rewrite. Compare also with StepAudio 3 Realtime Preview, or browse audio-to-audio models.
Compare the model on the work you need to complete.
Eleven Realtime v1 Mini is listed for Audio, Audio-to-audio. It accepts audio, text and produces audio, text. Use the publisher’s input format and preprocessing for your chosen deployment.
Compare pronunciation, voice consistency and time to the first playable audio. Check the service’s billing unit for the actual script.
Use the model identifier and account requirements for the selected API route. Confirm its request limits and output format with one complete task before connecting a larger workflow.
These records share a directory category. Check the source card, output type and deployment requirements before treating a model as a replacement.
rime
Rime Prism is a speech-to-speech model for enterprise voice agents that settles each reply as inspectable text first, with parallel tool calls.
Liquid AI
LFM3-Audio-2B is Liquid AI's open 2B speech model: ASR, TTS, voice chat and interleaved text-and-speech output from one local checkpoint.
Alibaba
Qwen-Audio-3.1-Realtime Flash is Alibaba's lighter full-duplex speech model with audio and text in and out, function calling, web search and a 262K context.
ElevenLabs
Eleven Realtime v1 Max is ElevenLabs' reasoning speech-to-speech agent for support and multi-step tool workflows, OpenAI Realtime-compatible.
xAI
Grok Voice 3.0 is xAI's speech-to-speech model for real-time voice agents, replacing Grok Voice 2.0 at the same $0.08-per-minute price.
Amazon
Nova 2.5 Sonic is Amazon's speech-to-speech model on Bedrock for real-time voice agents, with async tool calling, polyglot voices and a 256K context.