Linear per second, bounded by the model's own ceiling

    What does a 60-second AI talking head cost?

    Between 60 and 1,800 credits — exactly double a 30-second clip on any model that can take the job. What the doubling hides is that 2 of the catalogue's 19 priced lipsync rows publish a maximum credit ceiling below what 60 seconds needs, so the render isn't offered as a single job at any price. 17 rows go the distance; check the model before scripting a full minute.

    Billed per second of finished clipper_second

    The formula

    credits = rate_per_second × 60, only where rate × 60 ≤ the model's published ceiling
    rate_per_second
    the same 1-to-30-credit spread as any other lipsync length — a property of the model, not of the clip.
    the model's published ceiling
    maxCredits in the catalogue's own price matrix. Below it, 60 seconds is rate times 60 like any other length. At or above it, no premium unlocks the job — that duration simply isn't offered in one render.

    Worked example — A 60-second piece to camera on Sync Lipsync 2.0

    1. 1.rate_per_second = 5 credits
    2. 2.seconds = 60
    3. 3.published ceiling = 600 credits — comfortably above 5 × 60
    4. 4.5 × 60

    = 300 credits

    The question the 30-second page never has to ask

    Thirty seconds is an ad length. Sixty seconds is closer to what a talking head actually is on most platforms — a vlog cold open, a training clip, a full-length UGC read — and it is long enough that a question the shorter page never has to answer starts to matter: can the model you picked even render this in one job?

    Every one of the 19 priced lipsync rows clears 30 seconds without strain, which is why the shorter page never checks. At 60 seconds, 2 of them cannot — their own published maxCredits sits below rate × 60, so the render isn't just expensive, it isn't offered as a single job at all.

    The check costs nothing: the catalogue publishes the ceiling on every model page, and confirming a model clears a minute is one division. The expensive version of this mistake is scripting the full minute, generating it, and finding out afterwards.

    Every rate at 60 seconds, and which models get there

    Every figure below is computed from Versely's live price matrix — the same matrix the app quotes you from before it charges. These are prices, not verdicts: the buyer's guides are where models get ranked.

    Every per-second rate, run to 60 seconds

    The same spread as the 30-second page, doubled — plus whether the model's own ceiling actually allows it.

    Credits per secondExample model60s costFits in one job?
    1VEED Avatar Audio60Yes
    1VEED Avatars60Yes
    1VEED Lipsync60Yes
    4HeyGen Avatar V3240Yes
    5Sync Lipsync 2.0300Yes
    6Kling Avatar Standard360Yes
    8VEED Fabric 1.0 (SD)480Yes
    8VEED Fabric 1.0 Text (SD)480Yes
    9Sync Lipsync 2.0 Pro540Yes
    10Wan 2.2 Speech to Video (SD)600Yes
    10Wan 2.2 Speech Turbo (SD)600Yes
    10LTX 2.3 Audio to Video600No — exceeds published ceiling
    10HeyGen Avatar V5600Yes
    10HeyGen Image to Video600Yes
    10VEED Fabric 1.0 Fast (SD)600Yes
    12Kling Avatar Pro720Yes
    15VEED Fabric 1.0 (HD/4K)900Yes
    15VEED Fabric 1.0 Text (HD/4K)900Yes
    17Sync React 11,020No — exceeds published ceiling
    20Wan 2.2 Speech to Video (HD/4K)1,200Yes
    20Wan 2.2 Speech Turbo (HD/4K)1,200Yes
    20VEED Fabric 1.0 Fast (HD/4K)1,200Yes
    30Avatar X Reference to Video1,800Yes
    30Avatar X Text to Video1,800Yes

    2 of 24 rows are flagged. Every one of them clears 30 seconds, which is why this constraint is invisible on a shorter brief.

    The models that stop short of a minute

    Published ceiling ÷ rate is the longest single job the model's own price matrix allows — an arithmetic limit, not a policy.

    ModelRatePublished ceilingLongest single job
    LTX 2.3 Audio to Video10/sec20020s
    Sync React 117/sec25114s

    A script written for one of these models has to land under its own ceiling, or split across more than one generation — which is an editor assembly job, not this formula.

    Doubling the length at both ends of the rate spread

    Where the model supports it, the doubling is exact — no rounding, no base fee.

    Rate30s cost60s costDifference
    1/sec (cheapest)3060+30
    30/sec (priciest)9001,800+900

    The gap between cheapest and priciest stays 30× at either length — doubling the clip doesn't change which model is the expensive choice, it raises the stakes on having picked it.

    What changes a full-minute lipsync bill

    Whether the model can reach 60 seconds at all

    The lever this length adds that 30 seconds doesn't have. 2 rows are ruled out before price even enters the decision.

    Which model renders it

    Still the dominant cost lever among the rows that qualify — 60 against 1,800 for the same minute.

    Splitting a long script across two jobs

    The workaround on a model that stops short of a minute — two renders stitched in the editor, which is a separate export charge, not a discount on this formula.

    Audio length

    Still directly proportional — trimming the read to 50 seconds removes exactly a sixth of the bill.

    What a 60-second charge does not cover

    • The voiceover itself, billed separately per 1,000 characters of script.
    • Stitching two shorter renders into one continuous minute — that's an editor export, priced on its own page.
    • Captions burned onto the finished clip, their own render with their own preset multiplier.

    How many one-minute clips a month of credits buys

    A Monthly plan's 580 credits covers 9 one-minute clips at the 1-credit rate, or 0 at the 30-credit rate — roughly half the 30-second count either way.

    Plan sizes and what each grants are on the cost hub.

    Lipsync models checked against their own ceiling

    Every model page shows its own price matrix, its published credit range, and the option table this formula reads from.

    Frequently asked questions

    Is a 60-second talking head just double the 30-second price?+

    On a model that can take the job at all, yes — the formula is exactly linear. But 2 of the catalogue's 19 priced lipsync rows publish a ceiling below what 60 seconds needs, so on those the honest answer is "not available in one job", not a price at all.

    How do I know if my model can render a full minute?+

    Divide the model's published maximum credits by its per-second rate. If the result is under 60, that model needs the script split across more than one render.

    Why doesn't the 30-second page mention this ceiling?+

    Because nothing in the catalogue stops short of 30 seconds — every priced lipsync row clears it comfortably. The constraint only appears once the clip gets long enough to bump into it.

    What happens if I script something longer than the ceiling allows?+

    The job isn't priced higher — it's a length that model doesn't offer in a single pass. Splitting the script across two renders and joining them in the editor is the workaround, at the cost of a separate export charge.

    Does the language of the script change the price at 60 seconds?+

    No, exactly as at 30 seconds — nothing in this formula reads the language, only the duration and the model's own rate.

    Different job, different arithmetic

    These are not variations on the formula above — each one meters something else entirely.

    Which model should I actually pick?

    This page prices the job. The buyer's guides rank the models, with the sort criterion stated on every one.

    Jobs billed this way

    Further reading

    See the price before you spend it

    Versely quotes the exact credit cost of every generation before it runs, and the agent will estimate a whole batch on request. The all-in-one AI studio for creators. 60+ models for video, image, voice, music and lipsync in a single app.