Eleven Flash v2.5

Audio modelCard summarized

Our fast, affordable speech synthesis model

Source: Meet Flash (launch post)

Use it via API

$0.04 / 1K chars

Code example Compare

Where to call it

ElevenLabs APIeleven_flash_v2_5

Supports

Streaming

Full feature support

Good for

Real-time voice agents and chatbots (Agents Platform)
Interactive applications and games that need immediate responses
Large-scale bulk text-to-speech at lower cost

Overview

ElevenLabs · 04TTS

Flash v2.5

Our fast, affordable speech synthesis model

Typical price

$0.04 / 1K chars

Flash / Turbo text to speech

Max length: ~40 min (40,000 characters)
Languages: 32 languages (Multilingual v2's 29 plus Hungarian, Norwegian, Vietnamese)
Modes: text-to-speech, streaming
Released: 18 Dec 2024

Flash is ElevenLabs' low-latency TTS model, generating speech in about 75 ms plus application and network latency, and the recommended model for conversational voice agents. Flash v2 is English only; Flash v2.5 supports 32 languages. ElevenLabs says it trades slightly lower quality and emotional depth than Turbo for much lower latency. launch post

eleven_flash_v2_5

Typical price: $0.04 / 1K chars
Max length: ~40 min (40,000 characters)
Languages: 32 languages (Multilingual v2's 29 plus Hungarian, Norwegian, Vietnamese)
Weights: Closed
Released: 18 Dec 2024

text-to-speech
streaming
WebSocket input streaming
voice cloning (instant and professional)