![]()
The HeyGen Avatar 6 API is live on PoYo with two model IDs: heygen-avatar-6 and the Turbo tier heygen-avatar-6-turbo, at half the 1080p price. Both turn a photo, a reference video of an animated character, or one of HeyGen's 1,248 public avatar looks into a lip-synced talking MP4 at 1080p or 2K, driven by your audio or, for public avatars, by text. You can try both tiers on the HeyGen Avatar 6 model page before writing code.
This overview covers what HeyGen shipped, the request format on PoYo, the limits your code has to validate, the price per second, and what we measured on launch day.
What is HeyGen Avatar 6?
HeyGen Avatar 6 is HeyGen's talking-avatar video model. HeyGen's technical report, dated November 17, 2026, says "Avatar 6 reproduces how a person looks, speaks, and moves, up to 2K," and that it works "from a single photo." HeyGen describes two tiers, Avatar 6 Quality and Avatar 6 Turbo; on PoYo these are heygen-avatar-6 and heygen-avatar-6-turbo.
| Date | What happened |
|---|---|
| November 17, 2026 | HeyGen publishes the Avatar 6 technical report |
| November 19, 2026 | Both tiers callable on PoYo as heygen-avatar-6 and heygen-avatar-6-turbo |
Three claims from HeyGen matter to an integration, because they map to request fields:
- Native 2K. HeyGen says "every frame [is] generated at 2880×1620." On PoYo you choose it with
"resolution": "2K". - Motion prompts. HeyGen's report says: "Write the action next to the words. Avatar 6 performs it right on cue." On PoYo this is the
promptfield, up to 1,200 characters. - Preference over Avatar V. In HeyGen's online vote, 69.7% of 879 votes preferred Avatar 6 Quality over Avatar V, and 64.3% of 863 votes preferred Avatar 6 Turbo over Avatar V Turbo. These are HeyGen's own numbers.
HeyGen Avatar 6 API at a glance
| Item | heygen-avatar-6 | heygen-avatar-6-turbo |
|---|---|---|
| HeyGen tier | Avatar 6 Quality | Avatar 6 Turbo |
| Submit endpoint | POST https://api.poyo.ai/api/generate/submit | Same |
| Result | GET /api/generate/status/{task_id} or callback_url | Same |
| Input modes | Photo, reference video, public avatar | Same three |
| Output | One MP4, 1080p (default) or 2K | Same |
| Speech input | Audio 2–60 s, or text up to 1,000 characters (public avatar only) | Same |
| Price, 1080p | 20 credits ($0.10) per second | 10 credits ($0.05) per second |
| Price, 2K | 29 credits ($0.145) per second | 18 credits ($0.09) per second |
The request fields are identical, so switching tiers means changing only the model string. Field rules are in the PoYo HeyGen Avatar 6 docs.
Three input modes, one endpoint
There is no mode field. The inputs you send pick the mode, in this order:
| Mode | Selected when | Inputs | Aspect ratio |
|---|---|---|---|
| Public avatar | avatar is present | avatar + text (+ optional voice), or avatar + audio_url | aspect_ratio: 16:9 (default), 9:16, 1:1, 4:5, 5:4 |
| Reference video | video_url is present (and no avatar) | video_url + audio_url, optional one photo in image_urls | Follows the input video |
| Photo | Default | image_urls (exactly one) + audio_url | Follows the input photo |
prompt and resolution work in every mode. aspect_ratio and voice belong to public avatar mode only. Send the inputs of one mode per request. HeyGen's own developer docs split this work across two separate APIs (a model API for image or video plus audio, and an application API for HeyGen avatar looks); on PoYo it is one request shape for both.
Reference video mode is for animated or illustrated characters. For a real person, use photo mode with one photo.
Limits to validate before you submit
| Field | Rule |
|---|---|
image_urls | Exactly one URL; image_urls[0] is used |
video_url | At least 2 seconds, up to 50 MB |
audio_url | 2–60 seconds, up to 15 MB; the video runs the length of the audio |
text | 1–1,000 characters (public avatar mode, instead of audio_url) |
avatar | One of 1,248 names (156 people); letter case and extra spaces ignored |
voice | One of 102 names; default Warm Pro Narrator |
prompt | Up to 1,200 characters |
resolution | 1080p or 2K |
All media URLs must be public and directly downloadable over HTTP or HTTPS. PoYo checks the audio length, the reference video length, and the avatar name at submission, so bad inputs fail fast rather than after a wait. The output can be about a tenth of a second shorter than the audio. Every avatar and voice name is in Avatars and Voices, with a downloadable JSON catalog.
HeyGen Avatar 6 API pricing on PoYo
PoYo bills per second of output in credits (1 credit = $0.005):
| Model | 1080p per second | 2K per second | 10 s at 1080p | 60 s at 1080p |
|---|---|---|---|---|
heygen-avatar-6-turbo | 10 credits ($0.05) | 18 credits ($0.09) | 100 credits ($0.50) | 600 credits ($3.00) |
heygen-avatar-6 | 20 credits ($0.10) | 29 credits ($0.145) | 200 credits ($1.00) | 1,200 credits ($6.00) |
Billed seconds are rounded up to whole seconds. With audio_url, the charge follows the audio length and is taken at submission. With text, PoYo reserves credits when you submit and settles to the finished video's length when the task ends. The full breakdown, with worked examples, is in HeyGen Avatar 6 API pricing.
Launch-day test results
The PoYo team ran these on the production API on November 19, 2026, in public avatar mode with text. Wall time is submit to result as seen by a polling client, so read it as approximate.
| Model | Avatar / voice | Script | Output | Cost at current prices | Wall time |
|---|---|---|---|---|---|
heygen-avatar-6-turbo | Abigail Office Front / Jessica Anne Bogart | 27 words, 16:9, 1080p | 10.16 s | 110 credits (11 s × 10) | About 2.4 min |
heygen-avatar-6 | Bryan Tech Expert / Andrew | 21 words, 16:9, 1080p | 6.88 s | 140 credits (7 s × 20) | About 2 min |
Two minutes or more for a short clip means this is an asynchronous API: submit, keep the task_id, and poll or receive a callback. Do not hold a user-facing HTTP request open while it runs.
HeyGen Avatar 6 API quick start
This is the docs example for public avatar mode with text:
curl https://api.poyo.ai/api/generate/submit \
-H "Authorization: Bearer $POYO_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "heygen-avatar-6-turbo",
"input": {
"avatar": "Abigail Office Front",
"text": "Hello and welcome to this demonstration of HeyGen Avatar 6.",
"voice": "Warm Pro Narrator",
"prompt": "Speak with calm confidence and nod slightly at the end.",
"aspect_ratio": "16:9",
"resolution": "1080p"
}
}'
Check code in the body (it must be 200), save data.task_id, then call GET https://api.poyo.ai/api/generate/status/{task_id} until status is finished and read the MP4 from data.files[].file_url. Generated videos stay available for 3 days, so copy them to your own storage. Photo mode, reference video mode, a bounded Python loop, and error handling are in the HeyGen Avatar 6 API guide.
Which requests belong on the HeyGen Avatar 6 API
| You need | Call |
|---|---|
| A presenter video from a script, no audio file | heygen-avatar-6 or -turbo in public avatar mode (text-to-avatar walkthrough) |
| A real person's photo speaking your audio, 1080p or 2K | Photo mode |
| A mascot or cartoon character speaking | Reference video mode |
| Drafts, previews, high volume | heygen-avatar-6-turbo (see Avatar 6 vs Turbo) |
| The lowest per-second price from a photo plus audio | Compare with Kling Avatar 2.0 |
| Audio longer than 60 seconds | OmniHuman 2.0 takes 4–300 s of audio at 720p, 21 credits per second |
Frequently asked questions
What are the HeyGen Avatar 6 API model IDs?
heygen-avatar-6 (HeyGen's Avatar 6 Quality tier) and heygen-avatar-6-turbo. Both use POST https://api.poyo.ai/api/generate/submit with the same input fields.
How much does the HeyGen Avatar 6 API cost?
On PoYo, per output second: Turbo 10 credits ($0.05) at 1080p and 18 credits ($0.09) at 2K; Avatar 6 20 credits ($0.10) at 1080p and 29 credits ($0.145) at 2K. Seconds are rounded up.
Can the HeyGen Avatar 6 API make a video from text only?
Yes, in public avatar mode. Send avatar and up to 1,000 characters of text, plus an optional voice from the 102 listed voices. Photo and reference video modes need audio_url.
How long can a HeyGen Avatar 6 API video be?
With audio, 2–60 seconds, matching the audio. With text, the video runs as long as the avatar takes to read up to 1,000 characters.
What resolutions does the HeyGen Avatar 6 API support?
1080p (default) and 2K. Public avatar mode adds aspect_ratio (16:9, 9:16, 1:1, 4:5, 5:4); the other modes keep the input's aspect ratio.
Does the HeyGen Avatar 6 API support webhooks?
Yes. Add a top-level callback_url and PoYo posts the task result when it is finished or failed. You can also poll the status endpoint.