Introduction
Accelerated Avatar 6 inference tuned for quick turnaround and bulk personalized avatar workflows.
| Field | Value |
|---|---|
| Input | Image, Audio |
| Output | Video |
| Model identifier | heygen/avatar-6-turbo/talking-avatar |
| Engine family | HeyGen — Avatar 6 Turbo release |
| Primary media inputs | Portrait image (JPEG/PNG/WebP, max 32 MB) plus speech audio (MP3/WAV) |
| Audio length window | 1 second minimum up to 600 seconds (10 minutes) per clip |
| Resolution presets | 1080p (1080px short edge, default) or 2k (1620px short edge) |
| Custom frame sizing | Optional paired width and height, even integers from 256 to 3840 |
| Gesture and prompt tuning | Optional prompt (up to 1,200 chars) and expressiveness (default or less) |
| Reference video input | Optional single MP4/WebM clip (at least 2s; animated/illustrated on standard workspaces) |
| SpicyAPI per-second rates | $0.06/s at 1080p; $0.105/s at 2K (billed per whole input audio second) |
| Verified test output | 3.79s portrait audio at 1080p delivered a clean 1088×1632 MP4 |
Key features & highlights
Moody late-night delivery built for volume
Priced at $0.06 per second for 1080p output, Avatar 6 Turbo halves per-second spend relative to standard Avatar 6 while keeping script-matched head turns and facial nuance. In our lounge sample, a portrait of a man in an unbuttoned dark silk shirt under amber bar light speaks straight into the lens with an unhurried half-smile and natural nods. That balance makes high-volume localization and late-night editorial clips economical without sacrificing atmosphere.
Sharp low-key detail across darker studio frames
Switching resolution to 2k produces a 1620-pixel short edge for roughly 40% less than standard Avatar 6 ($0.105/s versus $0.174/s). Our dim control-room sample pairs a woman in an off-shoulder black knit top with a studio speech track, preserving clean eye-line focus, soft shadow gradients, and steady lip synchronization across the frame. Teams rendering multi-language tours or close-up host intros gain crisp 2K clarity on a batch-friendly budget.
Draft on Turbo, promote select hero cuts
Avatar 6 Turbo accepts the exact same portrait, audio, motion prompt, resolution, expressiveness, reference video, and custom dimension fields as standard Avatar 6. Product and localization pipelines can test scripts, compare vocal takes, and ship hundreds of regional clips on Turbo, then swap a single model identifier when a flagship hero cut warrants standard Avatar 6.
Request parameters
| Parameter | Required | Description |
|---|---|---|
image_url | Yes | HeyGen Avatar 6 Turbo: portrait photo or illustration (up to 32 MB). at least 1 characters |
audio_url | Yes | HeyGen Avatar 6 Turbo: speech audio driving lip sync and clip length. at least 1 characters |
prompt | No | HeyGen Avatar 6 Turbo: optional plain-text motion and gesture direction (up to 1,200 chars). at least 1 characters · at most 1200 characters |
resolution | No | HeyGen Avatar 6 Turbo: 1080p or 2k output tier (changes per-second rate). |
expressiveness | No | HeyGen Avatar 6 Turbo: default for natural gestures or less for calmer delivery. |
reference_video_urls | No | HeyGen Avatar 6 Turbo: optional single reference video (≥ 2s; illustrated/animated on standard workspaces). at least 1 items · at most 1 items |
width | No | HeyGen Avatar 6 Turbo: optional even output width in pixels (paired with height). |
height | No | HeyGen Avatar 6 Turbo: optional even output height in pixels (paired with width). |
Pricing
| Specification | Price |
|---|---|
| $0.06 / second |
| $0.105 / second |
Use cases
Multilingual product tours at scale
Feed one authorized presenter portrait alongside translated MP3 or WAV narration tracks to publish dozens of synchronized regional walkthroughs in parallel. At $0.06/s in 1080p, a 100-clip batch of 30-second localized tours totals $180 rather than $360.
Rapid script and performance prototyping
Compare vocal pacing, motion prompts, and calmer versus lively delivery before locking a final edit. Because failed jobs carry zero charge and 1080p runs cost six cents per audio second, iterating on short takes stays predictable.
Atmospheric late-night editorial hosts
Pair low-key portrait photography—amber lounges, candlelit balconies, or neon-washed studios—with intimate narration for entertainment recaps, fashion teasers, or podcast video promos that rely on poised eye contact rather than sterile studio lighting.
Long-form course and onboarding modules
Animate continuous spoken lessons up to 600 seconds in a single request, avoiding awkward mid-sentence splices. Context-driven movement keeps posture shifts aligned with the phrasing across multi-minute explanations.
Personalized campaign and CRM drops
Generate tailored account greetings, event invitations, or seasonal offers programmatically from a single approved host image and dynamically synthesized speech files.
Prompt guide & pro tips
Confirm portrait and voice rights
Select a sharp portrait (JPEG, PNG, or WebP up to 32 MB) and an MP3 or WAV recording of a consenting adult whom you hold full authorization to depict.
Check speech length and pacing
Keep
audio_urlbetween 1 and 600 seconds, since total charge scales by whole seconds of input audio rounded upward.Pick tier and performance direction
Pick 1080p ($0.06/s) for rapid batches or 2k ($0.105/s) for finer detail, then add an optional prompt up to 1,200 characters and choose default or less expressiveness.
Inspect framing and save output
Omit width and height to preserve your portrait aspect ratio, or supply an even pixel pair between 256 and 3840, then verify lip sync and archive the MP4.
Overview
heygen/avatar-6-turbo/talking-avatar animates a single portrait image and speech audio track into a synchronized 1080p or 2K MP4 video. As the speed- and budget-oriented sibling to HeyGen Avatar 6, Avatar 6 Turbo preserves script-aware facial motion, non-repeating gestures, and an identical API contract while cutting 1080p per-second spend in half ($0.06/s vs $0.12/s) and reducing 2K spend by roughly 40% ($0.105/s vs $0.174/s).
In our live verification on 2026-11-19, submitting a 3.79-second portrait speech track at 1080p produced a clean 1088×1632 MP4 with matched lip movement and preserved source framing.
Parameters
| Parameter | Required | Type / Values | Role |
|---|---|---|---|
image_url | Yes | URI (JPEG, PNG, WebP, max 32 MB) | Single-subject adult portrait or illustration; output inherits its aspect ratio unless overridden. |
audio_url | Yes | URI (MP3 or WAV, 1–600s) | Speech track driving lip sync, rhythm, and clip duration. Sets the billing basis. |
prompt | No | String (1–1,200 chars) | Plain-language cues for expression, eye contact, posture, and hand movement. Overrides expressiveness where overlapping. |
resolution | No | 1080p (default) or 2k | Output tier: 1080px short edge (1080p) or 1620px short edge (2k). Changes per-second rate. |
expressiveness | No | default or less | Gesture range: default for natural script-responsive motion, less for calmer, understated movement. |
reference_video_urls | No | Array of 1 URI (MP4/WebM, ≥ 2s) | Optional motion clip for animated or illustrated characters on standard workspaces. |
width | No | Even integer (256–3840) | Exact output width in pixels; must be paired with height. |
height | No | Even integer (256–3840) | Exact output height in pixels; must be paired with width. |
Pricing and Batch Cost Math
Billing follows input audio duration rounded up to whole seconds (1-second minimum, 600-second cap). Failed requests carry zero charge. Current SpicyAPI rates are $0.06 / second at 1080p and $0.105 / second at 2K (verify live pricing on the HeyGen Avatar 6 Turbo model page).
| Clip Length | 1080p Single Clip | 2K Single Clip | Batch Example (1080p) | Batch Example (2K) |
|---|---|---|---|---|
| 30 seconds | $1.80 | $3.15 | 100 localized tours = $180.00 | 100 localized tours = $315.00 |
| 180 seconds (3 min) | $10.80 | $18.90 | 20 course lessons = $216.00 | 20 course lessons = $378.00 |
| 600 seconds (10 min) | $36.00 | $63.00 | 10 full briefings = $360.00 | 10 full briefings = $630.00 |
Choosing Between Avatar 6 Turbo, Avatar 6, and Avatar IV
| Capability | Avatar 6 Turbo | Avatar 6 | Avatar IV |
|---|---|---|---|
| 1080p rate | $0.06 / s | $0.12 / s | Flat rate across 360p–1080p |
| 2K output | Yes ($0.105 / s) | Yes ($0.174 / s) | No (max 1080p) |
| Max audio length | 600 seconds | 600 seconds | 600 seconds |
| Motion controls | prompt + default/less | prompt + default/less | prompt + low/medium/high |
| Background swap | No (set scene in portrait) | No (set scene in portrait) | Yes (background_image_url) |
| Custom even pixel sizing | Yes (width + height) | Yes (width + height) | No (aspect_ratio presets) |
| Typical fit | Bulk localization, rapid drafts, daily runs | Flagship close-ups, unprompted acting nuance | Background replacement or sub-1080p tiers |
Workflow and Prompting Tips
Pair Turbo with concise motion cues
Community testing shows Avatar 6 Turbo tracks standard Avatar 6 closely when prompt states concrete gaze, posture, and mood cues—for example, leaning slightly toward the camera with a warm half-smile in a dim bar frame, or maintaining calm studio posture under neon lighting. Leave lighting and wardrobe to image_url, and reserve prompt for delivery.
Draft on Turbo before promoting hero cuts
Because both Avatar 6 models validate the same payload, run multi-language audio checks and script revisions on heygen/avatar-6-turbo/talking-avatar at 1080p. When a hero launch cut calls for standard Avatar 6, change the model identifier without rewriting field mappings.
Control audio and frame geometry upfront
Trim dead air from MP3 or WAV files before upload so you do not pay for silent tails. If you pass width and height, supply even integers that match your portrait orientation to prevent awkward cropping.
Consent, Rights, and Usage Boundaries
Submit portraits and voice recordings strictly of consenting adults whom you are authorized to represent. Confirm permission for both the visual likeness and the spoken track prior to calling the endpoint, and never use this model to impersonate anyone. Realistic human clips passed via reference_video_urls are blocked on standard workspaces; use an authorized still photo in image_url instead.
API quick start
# Generate the idempotency key once; reuse it when retrying after a lost response.
export SPICY_IDEMPOTENCY_KEY="${SPICY_IDEMPOTENCY_KEY:-$(uuidgen)}"
curl -X POST https://api.spicyapi.ai/api/v1/jobs/createTask \
-H "Authorization: Bearer $SPICY_API_KEY" \
-H "Content-Type: application/json" \
-H "Idempotency-Key: $SPICY_IDEMPOTENCY_KEY" \
-d '{
"model": "heygen/avatar-6-turbo/talking-avatar",
"input": {
"audio_url": "https://example.com/reference.mp3",
"expressiveness": "default",
"image_url": "https://example.com/reference.png",
"resolution": "1080p"
}
}'
# Copy data.taskId from the response into TASK_ID, then poll the status.
TASK_ID="TASK_ID_FROM_RESPONSE"
curl "https://api.spicyapi.ai/api/v1/jobs/recordInfo?taskId=$TASK_ID" \
-H "Authorization: Bearer $SPICY_API_KEY"
See the API reference for authentication, the complete schema, polling, and error handling.



