One currency, priced per job
Credits are the only unit Versely bills generation in. Each of the 385 models carries its own credit cost, and the exact figure for the settings you have chosen is shown before you confirm a generation — there is no post-hoc metering and no surprise on the invoice.
Two things decide what a job costs: the model, and how much work you have asked it to do. Some models are flat — the same price whatever you type. Most video and lipsync models are not, because they are billed on what they render: seconds of output, resolution, whether a native audio pass runs.
The shape of the catalog
Where the money goes, by format. “Fixed price” counts the models that charge the same whatever you ask for; the rest quote a band, because they bill on what they actually render. Floor and ceiling are the cheapest and dearest single generation possible in that format.
| Format | Priced per generation | …of which fixed price | Cheapest generation | Dearest generation |
|---|---|---|---|---|
| Video | 197 | 2 | 1 cr | 1,988 cr |
| Image | 138 | 119 | 1 cr | 28 cr |
| Speech & audio | 6 | 5 | 1 cr | 480 cr |
| Lipsync & avatars | 23 | 1 | 2 cr | 2,880 cr |
| Video upscale | 1 | 0 | 8 cr | 192 cr |
20 speech models are excluded from this table because they are not billed per generation at all — they bill by the length of the script, per 1,000 characters. The floor on a banded model is the cheapest legal combination of settings and the ceiling is the most expensive one: a short 480p clip and a long 4K clip on the same model are genuinely different jobs, which is why one number would be a lie in either direction.
What the featured models cost
The 21 models flagged as featured in the catalog that also have a page of their own — roughly what the app puts in front of you first — with their credit cost and what moves it. This is a slice, not the whole catalog: 206 models publish the same figures on their own model page.
| Model | Format | Credits per job | What moves it |
|---|---|---|---|
| PrunaAI P-VideoPrunaAI | Video | 2–32 cr | Length of the output |
| Qwen Image 3 Text to ImageQwen | Image | 3 cr | Output megapixels |
| VEED AvatarsVEED | Lipsync & avatars | 3–56 cr | Length of the output |
| GPT Image 2.5 Flare EditOpenAI | Image | 4–7 cr | quality resolution flat |
| Wan 3.0 Text to VideoWan | Video | 4–120 cr | Resolution × length |
| Seedream 5.0 ProByteDance | Image | 6 cr | Nothing — one fixed price |
| Midjourney V7Midjourney | Image | 8 cr | Nothing — one fixed price |
| Wan 2.7 Image to VideoWan | Video | 8–90 cr | Resolution × length |
| Wan 2.7 Text to VideoWan | Video | 8–90 cr | Resolution × length |
| Minimax H3 Text to VideoMiniMax | Video | 12–78 cr | Resolution × length |
| HeyGen Avatar V3HeyGen | Lipsync & avatars | 14–327 cr | Length of the output |
| Happy Horse 1.1 Image to VideoAlibaba | Video | 17–108 cr | Resolution × length |
| Happy Horse 1.1 Text to VideoAlibaba | Video | 17–108 cr | Resolution × length |
| Seedance 2.0ByteDance | Video | 22–1,639 cr | Resolution × length |
| Seedance 2.5ByteDance | Video | 22–320 cr | Resolution × length |
| Flux 3 Text to VideoBlack Forest Labs | Video | 34–136 cr | Length of the output |
| HeyGen Avatar V5HeyGen | Lipsync & avatars | 40–960 cr | Length of the output |
| HeyGen Image to VideoHeyGen | Lipsync & avatars | 40–960 cr | Length of the output |
| Sync React 1Sync | Lipsync & avatars | 67–201 cr | Length of the output |
| Avatar X Text to VideoMirage | Lipsync & avatars | 120–2,880 cr | Length of the output |
| Runway Gen-4.5Runway | Video | 200–400 cr | Length of the output |
Hunting for the cheapest option in a format rather than the featured one? The ranked guides sort the whole catalog by credits.
What a finished video costs
Single generations are the wrong unit for planning. These are the 17 published workflow recipes — real multi-scene videos with real scene counts — and the credit total each one comes to. Where a workflow lists two figures, the first is the preview render and the second the full-resolution one.
| Project | Scenes | Credits | Video model |
|---|---|---|---|
| Viral Panel & Reaction Reel | 4 | 32 / 96 cr | Seedance v1.5 Pro Image to Video |
| Podcast with me | 4 | 92 / 248 cr | Vidu Q3 Image to Video |
| Podcast Clip — Versely | 2 | 108 / 216 cr | Kling Video V3 Pro Image to Video |
| Plush Lipstick Influencer Review | 3 | 108 / 297 cr | Vidu Q3 Image to Video |
| Morning Matcha Routine | 3 | 162 / 324 cr | Kling Video V3 Pro Image to Video |
| Hivewell Raw Honey | 10 | 180 / 500 cr | Vidu Q3 Image to Video |
| Digestive System Explainer | 11 | 198 / 396 cr | Kling 3 Turbo Image to Video |
| NYC Street Interview | 3 | 204 / 405 cr | Kling Video V3 Pro Image to Video |
| The Sweet Switch: Fruit Heist | 7 | 259 / 700 cr | VEO 3.1 Fast I2V |
| AlignPro — Posture Corrector Ad | 15 | 262 / 724 cr | Vidu Q3 Image to Video |
| Cat-Faced Cabin Heist | 8 | 288 / 792 cr | Vidu Q3 Image to Video |
| NoseFresh — Nasal Strips Ad | 15 | 345 / 930 cr | Vidu Q3 Image to Video |
| The Secret of the Golden Grain | 6 | 360 / 960 cr | VEO 3.1 Fast I2V |
| NEURA — Coffee vs Brain | 18 | 414 / 1,116 cr | Vidu Q3 Image to Video |
| LUMÉE GLOW — Rich Cat Glow-Up Story | 28 | 644 / 1,736 cr | Vidu Q3 Image to Video |
| Primo Protein vs Other Brand | 11 | 660 / 1,760 cr | VEO 3.1 Fast I2V |
| Lunchwell — Hospital Visit Story | 18 | 1,080 / 2,880 cr | VEO 3.1 Fast I2V |
How far each plan goes
Each cell is one cycle’s credits divided by that job’s cost, rounded down. It assumes you spend the entire allowance on that single job type, which nobody does — treat it as the ceiling for one activity, not a forecast.
| Job | Weekly240 cr / week | Standard Monthly774 cr / month | Pro Monthly1,600 cr / month | Standard Annual9,600 cr / year | Pro Annual19,200 cr / year |
|---|---|---|---|---|---|
| One ERNIE Image Turbo imageCheapest fixed-price image model · 1 cr | 240× | 774× | 1,600× | 9,600× | 19,200× |
| One Higgsfield imageDearest fixed-price image model · 19 cr | 12× | 40× | 84× | 505× | 1,010× |
| The "Viral Panel & Reaction Reel" workflow4-scene video, preview render · 32 cr | 7× | 24× | 50× | 300× | 600× |
| The "NEURA — Coffee vs Brain" workflow18-scene video, preview render · 414 cr | 0× | 1× | 3× | 23× | 46× |
A 0 means one cycle’s credits do not cover a single run of that job — you would need to bank more than one cycle, or top up with a pack. Only jobs with a single unambiguous credit cost are used here. Variable-priced video models are deliberately excluded: dividing a plan by a model’s floor would flatter the answer, and dividing by its ceiling would scare you off. Their bands are in the table above instead.
Unused credits roll over
Renewal grants are additive: each cycle adds its credits to whatever is already in your balance instead of resetting it. A quiet month is not a wasted one for as long as the subscription stays active. One-time packs top up that same balance rather than sitting in a second wallet.
Nothing generates for free
The cheapest model in the catalog is 0 credits and no model prices at zero, so there is no free plan and no daily free allowance. New accounts start with no credits; the only credits outside a plan or a pack are a first week's: 100 credits (80 + 20 bonus) with the $3 first week on the web (on the Weekly plan, for the first 100 customers), or the same with the iPhone app's $2.99 first week. Both are on the pricing page. The smallest entry point is the cheapest weekly plan; once subscribed, the $1 pack adds 27 credits on top.
Credit questions
What is one Versely credit?+
A credit is the single unit Versely bills generation in. Every model publishes its credit cost, and the exact figure is shown before you confirm a generation. Across the 385 models in the catalog, one generation runs from 0 credits at the low end to 2,880 for the longest, highest-resolution video a model will render.
Why do two videos on the same model cost different amounts?+
Because most video and lipsync models are billed by what they actually render. Length, output resolution, and whether a native audio pass runs all move the number, which is why the table above gives a floor and a ceiling for each model rather than one price.
What does a finished video cost, start to finish?+
The published workflow recipes are the honest answer, because each one is a real multi-scene video with a real total. They run from 32 credits for "Viral Panel & Reaction Reel" (4 scenes) up to 1,080 credits for "Lunchwell — Hospital Visit Story" (18 scenes) at preview resolution.
What is the cheapest thing I can generate?+
0 credits, which is the floor across the whole catalog. Nothing generates for zero credits and there is no daily free allowance. New accounts start with no credits; the only credits that do not come from a plan or a pack are the trial's ($3 for your first week on the web or $2.99 on iPhone, with 100 credits (80 + 20 bonus)).
Do credits expire?+
Not while you are subscribed. Each renewal adds its credits to your existing balance instead of resetting it, so an unused month carries forward.
Can I see the cost before I generate?+
Yes. Every generation screen shows the credit cost for the exact settings you have chosen — model, length, resolution — before you confirm it.
Pick a plan against your real numbers
Now you know what your work costs in credits, the plan table is a one-line decision. The all-in-one AI studio for creators. 60+ models for video, image, voice, music and lipsync, in your browser or on your phone.