Tasks and inputs
HeyGen Voice is listed for Audio, Text → Speech, Text-to-speech. It accepts text and produces audio. Use the publisher’s input format and preprocessing for your chosen deployment.
HeyGen Voice is HeyGen's default text-to-speech model: about 39 languages, public and cloned voices, streaming with character timestamps.
The specification score summarizes recorded features and limits. It is not a benchmark result.
Pricing conditions: Indicative and conflicting: $15 per 1M characters via OpenRouter vs $30 per 1M characters on HeyGen's own API pricing. Confirm on your account.
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
Most text-to-speech APIs make you choose between stock voices and your own. HeyGen Voice handles both in one model. It is HeyGen's default TTS engine, and it speaks with public library voices, instant clones made from a short sample, or professional clones built with more care.
The professional voices get the finer controls. You can set a seed for repeatable output, adjust speed and pitch, and insert break tags to place pauses exactly. Instant clones and public voices work, but without that full control surface.
Around 39 languages are covered, addressed by ISO 639-1 codes. Leave the code out and the model detects the language itself, which is handy for user-generated text but worth overriding when you already know the locale.
The ulaw and alaw options are the telephony giveaway. If you are piping audio into a phone system, you can skip a transcoding step. Character timestamps let you drive captions or lip-sync without a separate alignment pass.
This is the part to check carefully. Through OpenRouter, HeyGen Voice is listed at $15 per 1M characters. HeyGen's own voiceover pricing shows $30 per 1M characters for its API. That is a two-fold gap for what appears to be the same model. Before you forecast costs, price a real job on the account you will actually bill against.
The release date is also soft. We use 13 October 2026, but some coverage gives 9 October.
HeyGen also ships HeyGen Voice Turbo, a lower-latency variant that drops professional clones. Use this standard model when the voice itself is the product: a narrator, a brand voice, a cloned presenter. Use Turbo when the first audio byte has to arrive fast and stock or instant voices are good enough.
Write ten lines that stress the engine: a number-heavy sentence, a brand name, a question, an aside in brackets, and a line in your second language. Render each with a public voice and with a clone. Listen for pacing and mispronunciations, and time the first streamed chunk. Then compare against other text-to-speech models such as Gemini 3.8 Flash TTS.
Compare the model on the work you need to complete.
HeyGen Voice is listed for Audio, Text → Speech, Text-to-speech. It accepts text and produces audio. 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.
zyphra
ZONOS3-Mini is Zyphra's open multilingual text-to-speech model: 800M MoE with ~100M active, zero-shot voice cloning in 43 languages and 80 ms first audio.
Alibaba
Qwen-Audio-3.1-TTS Plus is Alibaba's premium text-to-speech model for brand voices and dubbing, with fine-grained emotion tags and noise-robust voice cloning.
Mistral AI
Voxtral TTS 3 is Mistral's speech synthesis model for batch and streaming audio, with voice cloning, emotion tags and 35+ languages at $13 per 1M chars.
fish-audio
Fish Audio Drama 3 is an expressive 35B MoE text-to-speech model steered by bracketed directions, covering 83 languages, with downloadable weights.
heygen
HeyGen Voice Turbo is the low-latency version of HeyGen's TTS model for public voices and instant clones, at $4 per 1M characters.
Alibaba
Qwen-Audio-3.1-TTS Flash is Alibaba's streaming text-to-speech model for voice assistants, live chat and customer service, with system and cloned voices.