HeyGen Pricing in 2026: What Your Plan Actually Costs Per Minute

The Free tier is region-dependent. The pricing card promises three videos a month. The help centre says "1 to 3 free videos, depending on your region." Both are first-party and they disagree in scope, so budget the lower number.

Extra Business seats do not buy credits. The card reads "Additional seats are $20 per seat." The help centre is blunter: "Purchasing additional seats does not increase your credits. The credit pool is shared between all members of the space." That sentence is the most expensive detail on this page, and I price it below.

Every Pro tier carries the same features. The vendor's own FAQ: "All Pro tiers include the same features: only the monthly credit allocation changes." The rungs run from 1,000 credits at $49 to 100,000 at $4,300, and the intermediate rungs are not published, so do not let a spreadsheet assume one.

Rollover is real and asymmetric. Unused credits roll over for one more month on monthly billing and accumulate until renewal on annual, and they do not survive cancellation. Read that middle clause carefully: you pay $288 up front on Creator annual and the credits still arrive monthly, not as a block.

Enterprise has no standard self-serve credit tier.

Two add-ons sit on top, and the vendor contradicts itself on one of them. Video Avatar Slots are listed at $29/month or $288/year in the current help article and at $29/month or $300/year in a legacy one. LiveAvatar Slots run $49/month on the LiveAvatar platform only.

Now the discount, computed rather than quoted. Creator at $288 a year is 17.2% off $29 monthly. Business seat 1 at $1,428 is 20.1% off $149. Additional Business seats carry no annual discount at all, at $240 a year against $20 monthly. A 20-seat Business workspace on annual billing earns its discount on one seat in twenty.

What a HeyGen AI video generator actually bills you for

Here is the finding that should reorganise how you read the page above. HeyGen does not price video by the minute in its plans. It prices it by the credit, and the credit's value changes with the engine you pick. The rate card lives in the help centre, not on the pricing page.

Source: help.heygen.com articles 15125761 and 9204682, both read 2026-09-29. The second covers the retired Premium Credit pool and states those rates "are now aligned with HeyGen's credit-based plans."

That table, not the plan ladder, is the vendor's real price list. Three consequences follow, and none of them appears on any page ranking for this query.

The "1 credit = 1 minute" premise is false, and the spread is 12x. On the same Creator plan, 600 credits buy 150 minutes of Avatar III Video or 12.5 minutes of Avatar V. Same money, same allowance, twelve times the output. Whoever told you a credit is a minute has never rendered Avatar V.

Short generative clips are the most expensive thing on the meter. A VEO3 video asset costs 90 credits regardless of length, so a ten-second clip and a two-minute clip bill identically, and that ten seconds is 15% of a Creator month. One minute of Video Agent in Seedance mode costs 120 credits, a fifth of the same month. If your team experiments with short generative clips rather than avatar presenters, you are on the worst-value line in the table.

Upscale is the most expensive line per minute. At 2 credits per second for 4K, one minute of upscaling costs 120 credits: 2.5 times Avatar V and nearly four times Avatar IV Video. Overage on this platform is not only a generation problem. It is an export problem too, which is the same credit-meter ambiguity we unpicked in Manus AI pricing.

The credit-overage worked example, and what the next minute costs

Overage here is tier-locked, and that changes the answer completely. Business customers can buy extra credits at $0.05 per credit, with a $5 minimum for 100 credits and $5 increments after that. Creator cannot buy credits at any price: "Creator does not support one-time credit purchases. If you need more credits, you can upgrade to Pro." Pro is the same shape one tier up, and auto-reload is Business-only. So on the two cheapest plans there is no marginal price to calculate, only an upgrade.

Now the worked example, on the engine HeyGen markets hardest. Avatar IV Video Look, 31 credits per minute, a target of 30 finished minutes a month. That workload is 930 credits.

Read the first row twice. A buyer who budgets a $29 tool and picks the engine the vendor advertises is looking at a $49 plan, 69% above what they budgeted, and the remedy is not a top-up, it is a plan change. The same 30 minutes on Avatar IV Photo Look at 16 credits a minute costs $0.97 per finished minute on Creator, a 41% saving for a different render mode on the same avatar. On Avatar III Video at 4 credits a minute, the same $29 covers it with room to spare.

Now the number nobody prints, the marginal minute:

Here is the reframe. Avatar IV Video's marginal minute on Business is $1.55. Creator's average minute for the same engine is $1.50. They sit within 3% of each other, so on this vendor overage is not a penalty tier, it is roughly the going rate. In one place the marginal rate beats the plan outright: Business credits work out at $0.0993 each at one seat, double the $0.05 top-up price. Buying top-ups on Business is cheaper per credit than buying Business.

That is a DERIVED figure. Take the vendor's price, divide by the vendor's allowance, and you get the vendor's own effective rate. The vendor publishes both halves and never performs the division. It is the same arithmetic we ran on Clay's two-meter credit model.

One assumption sits under every figure in this section, and it is aggressive: finished minutes equal generated minutes. The vendor publishes no re-render rate, and the review record says a re-render costs full credits again. One Trustpilot reviewer, quoted by checkthat.ai on 2026-07-10: "they use a credit system that I don't like because you only get a limited amount and each time you edit something, it uses credits, even though their pricing says unlimited videos." If a script needs two takes, halve every minute figure above and double every per-minute cost.

Cost at volume: solo, 5 seats, 20 seats

Three team shapes, one engine, and every figure derived from the vendor's own prices and rates. The assumptions are stated per scenario because they are the whole argument: Avatar IV Video Look at 31 credits per minute, first-try renders, finished minutes equal generated minutes.

Scenario A: a solo creator producing 30 finished minutes a month.

Scenario B: a 5-seat marketing team producing 60 finished minutes a month. That is 1,860 credits.

Scenario C: a 20-seat team producing 240 finished minutes a month. That is 7,440 credits.

A Pro tier sized to 7,500 credits is not priced here, because the intermediate Pro rungs are not published. Two tiers are verified: 1,000 credits at $49 and 100,000 at $4,300.

The pattern repeats at both team sizes, and it widens with headcount.

The mechanism is one number. Every extra Business seat costs $20 and brings zero credits. Seat 1 prices 1,500 credits at $0.0993 each. Seat 20 takes the same 1,500 credits to $0.3527 each, a 3.55x degradation for an identical allowance. The extra seat looks cheap next to a $29 Creator seat, and it is the expensive one, because the Creator seat arrives with 600 credits and the incremental Business seat arrives with none.

Now the counter-argument, because without it this reads as an unfair hit piece. Separate Creator seats buy none of the workspace layer: no shared credit pool, no shared asset library or brand kit, no 60-minute videos, no SAML or SSO, no SCORM export, no LMS integration, no centralized billing, no role or content access controls, no audit log, no video draft collaboration, and one custom video avatar per account instead of five. The Business premium is a governance premium, not a capacity premium. If your binding constraint is video volume, the seat is the wrong purchase. If it is compliance, asset control and a single bill, the seat is the right one and its sticker price is incidental. That seat-versus-usage split is the same one we unpicked in Notion pricing.

Creator or Business: the decision matrix

Forget the feature grid. If you are sizing a HeyGen video generator seat for a team, this decision turns on monthly volume and on whether anyone outside your team will audit the account. Every figure below is as of 2026-09-29.

If you are a solo HeyGen AI video creator choosing between Creator and Pro, the question is not which plan has more features. Every Pro tier has the same features. The question is whether your month exceeds 600 credits at the engine you actually render with, and on Avatar IV Video that ceiling arrives at 19.4 finished minutes.

The discount and promo layer

The vendor's help centre confirms a promo field exists: "If you've received a promo or discount code, you can apply it at checkout. Look for the Add promotion code field under the subtotal before you complete your purchase. We occasionally share codes like these, so be sure to follow us on social media to catch future deals." That is the entire first-party position. There is no published code list.

Third-party claims are unverified and they contradict each other. One coupon site advertises 20% off in its own page title, then lists codes at 10 to 20%, at 25%, and elsewhere claims 10 to 75%. Another advertises 50% off for three months and a "$1,000 for 20,000 credits" bundle gated behind a paid membership. Read all of it as affiliate inventory rather than vendor pricing.

The defensible verdict: the annual commit beats any code you can actually verify. The annual discount is vendor-published and it is arithmetic, at $24 against $29 on Creator and $119 against $149 on Business seat 1, and it recurs at every renewal while a first-purchase code does not. Carry one caveat. On Business the discount applies to seat 1 only, so on a 20-seat annual workspace you are saving on 5% of the seats. Do not let a 20% headline sell you a 20-seat commitment.

Where HeyGen pricing stops making sense

Three boundaries, each one honest about what it costs you to ignore it.

When no avatar is needed. Professionally produced corporate video runs $1,000 to $5,000 per finished minute, and remote production teams run $250 to $500. HeyGen at Avatar IV Video on Creator is $1.50 per finished minute, cheaper by between 167x and 667x. I would not spend that saving twice. The product makes a talking presenter. If the video needs to explain something rather than present something, the per-minute comparison is not the relevant one. The category-wide counterweight matters too: AI video tools deliver only 1.1x to 1.6x ROI, the weakest category in the 2026 data, "largely because production overhead remains high even when generation is automated" (Digital Applied, 2026, corroborated by Biziq). Both publishers add that AI-generated paid social creative underperforms, because Meta, TikTok and Google down-rank obvious AI creative in their 2026 ranking updates. The saving is real. The return is not proportional, because the remaining overhead is the script, the review cycle, the re-render and the distribution.

When the volume is low enough for Free. Free covers 1 to 3 videos a month at 1 minute, 720p and watermarked, at $0. The step from $0 to $29 is the largest relative jump on the ladder. If you produce under three minutes a month and can accept a watermark, you are paying for nothing.

When short generative clips eat the month. VEO3 at 90 credits per asset and Video Agent Seedance at 120 credits per minute are priced against runtime they do not have. Keep script-in-flux work on Avatar III Video at 4 credits per minute and move to Avatar IV or V only once the script is locked. A re-render at 31 or 48 credits per minute is the retake tax behind the most repeated complaint in the review record.

One honesty note forces the price question onto the vendor itself. checkthat.ai records a February 2026 HeyGen statement: "Many of you have told us the feature and credit system is confusing... In our own experiments, we validated that confusion." The same source notes that April 2026 reviews still cited the same confusion, and that the vendor responded with premium-feature labelling, pre-generation cost estimates and unlimited audio dubbing. G2 sits at 4.8/5 across roughly 1,475 to 1,800 reviews. Trustpilot is genuinely contested: four sources report 2.3, 2.4, 3.2 and 3.6 out of 5, while Trustpilot's own page headline reads "rated Great with 4/5". I am not printing one Trustpilot number as fact, because the evidence does not support one. A Reddit thread sits at position 14 on this query with a user reporting "HeyGen Avatar V works out to $370/mo" for 360 minutes a month, which does not reconcile with the published rate card: 360 minutes of Avatar V is 17,280 credits, or $864 at the only published marginal rate. Treat it as a user report, not a benchmark.

How to read these numbers

Measured, first-party, read 2026-09-29: every plan price, seat price, annual figure and credit allocation in the ladder table, every rate in the credit table, the $0.05 per-credit price and its Business-only availability, the rollover terms, the Creator and Pro top-up restrictions, the add-on prices and the promo-code field. Also measured, dated 2026-09-17: the vendor blog post stating Avatar IV at 20 credits per minute and translation at 2 and 5 credits per minute.

Derived by me from those inputs: every cost per finished minute, every effective price per credit, the 3.55x seat degradation, the marginal-minute table, the three scenario totals and the 167x to 667x production comparison.

Assumed, and left visible in the tables: the monthly volumes (30, 60 and 240 minutes), Avatar IV Video Look as the reference engine, and first-try renders. Change the first two to your own numbers and the arithmetic still holds. Change the third and every minute figure halves.

Unverified, therefore absent: the "$2 per credit" and "$1.00 per minute" figures circulating in explainers, any single Trustpilot score, every third-party promo code, the third-party API rates we treat the same way we do on how API pricing is metered, and the intermediate Pro tier prices between 1,000 and 100,000 credits. This is the same dated-rate discipline we hold in Grok pricing.

Bottom line: what I would buy at each team size

One recommendation per shape, then the part about video that nobody selling video software wants to discuss.

• Solo, under 20 finished minutes a month, Avatar IV Video: Creator at $29, and be ready to move to Pro at $49 the month the workload crosses 19.4 minutes.
• Solo or small team, 20 to 32 minutes a month: Pro base at $49, the only route that adds credits without adding the governance layer.
• Five operators, 60 minutes a month, no compliance requirement: five Creator seats at $145 for 3,000 credits beats one Business workspace at $249 for 1,900. That is 42% cheaper and 58% more credits.
• Any team where SSO, audit logging or a shared pool is the constraint: Business at $149 plus $20 per seat, with top-ups budgeted separately at $0.05 per credit. Do not buy Business for capacity, buy it for governance, and be honest that you are paying a premium for it.
• Anyone producing under three minutes a month: stay on Free and spend the $29 on the script.

The last thing worth saying is the thing I would say to my own team before we add a video line to a budget. Video is one channel in a campaign, not a separate subscription. The script, the review cycle, the ad variants and the reporting are the same work whichever tool renders the frames, and the ROI data above says the overhead travels with the workflow rather than with the render. That is the argument for running video inside the platform that already holds your campaigns and analytics rather than beside it. Allable covers that chain from one conversation, with the campaign, SEO, social and image work sitting in the same place as the video brief, and every plan starts with a 7-day free trial. Cost the video work as one line in a stack you already pay for, rather than as another seat, and see whether the arithmetic survives. The rest of your marketing stack usually does not, which is why workflow automation and AI marketing tools end up on the same budget review.