2026-09-29 · VideoFreeTier

Kling AI Free, Kling 3.0 Free and Kling Free Credits: Who Said the Number Matters More Than the Number

Kling's free tier is a trial package, not a daily allowance: 66 credits, and on the account observed here they carry an expiry date of 3 November 2026. That number comes from inside a logged-in session, read on 2026-10-03 — the public pricing URL does not carry it. Without a login it redirects to the marketing site, and a headed browser that read 36,382 characters of that page found zero credit or plan figures — re-checked on 2026-11-04, unchanged. The plan table itself, though, sits on a different URL that was found rendering with no login at all on 4 November 2026; the update below narrows the old claim by URL. One generation at 720P costs 9 credits per second, measured the same day, so the package funds about seven seconds of 720P video in total — and a single 10-second clip, at 90 credits, costs more than the entire grant. Kling 3.0 free, Kling 2.5 free and Kling free credits are one question about one grant at account level, and until late October this site held no version-qualified reading for Kling: the 9-credits-per-second figure above names no version. That changed on 26 October 2026, when a logged-in session priced Kling 4.0 specifically at 10 credits per second — the two rates, and what the newer one does to the seven-second estimate, are both below.

Watch the same finding as a 99-second video: youtu.be/fIHqskt06bM

Nothing on this site requires a sign-up. Every dated figure we hold is published as CSV, JSON and Markdown on the data page — no account, no email gate, nothing to register for before you can read a number.

Update, 4 November 2026: one URL still hides the numbers, another hands them over

The claim that the plan table renders only inside a logged-in session needs narrowing, and the narrowing is the finding. Re-checked on 4 November 2026 with a headed browser, a fresh browser profile and no cookies: https://kling.ai/pricing still redirects to the marketing homepage — 36,389 characters, zero plan or credit figures, same as on 3 October. But https://app.klingai.com/global/membership, which lands on https://kling.ai/app/membership/membership-plan, rendered the full plan cards and the plan-versus-plan comparison table with no sign-in at all; the only obstruction was a dismissible Kling AI Trial Package overlay inviting the reader to sign in, and it does not cover the table. The correction is by URL, not by site: the pricing path gives nothing, the membership path gives everything, and any sentence covering both with one login-wall claim was wrong as of this date. Whether the open rendering is deliberate or a temporary state is not something one observation on one date can answer.

What the free column does not include

With the table readable without an account, the free column can be quoted as published. The source is Kling's own comparison table, read on 4 November 2026, column Basic ($0 / Free), copied cell for cell:

Read that way, the only row carrying a daily allowance that the free column also receives is Element AI Multi-Shot at three free uses a day — a specific feature, not a credit balance. A dash in this table means one thing and no more: the benefit is not listed in the free column. It is tempting to read the watermark row's dash as the free tier adds a watermark, but no sentence on the page says that, so this page does not either. One row name deserves its own note: Daily Free Credits for Subscribers implies a number, and the free column's dash says free accounts sit outside it — but the row's own hover text, read the same day, carried only an offer-valid-until date of Dec. 31, 2026 and a platforms-only note, with no figure for what subscribers receive either.

Kling 4.0 rates, read inside a logged-in session on 26 October 2026

The 9-credits-per-second figure in the opening paragraph was read on 3 October 2026, and the metric it was filed under names no Kling version. The 4.0-specific check, done on 26 October 2026 from the generator's own cost line while logged in on a Free Trial Package account holding 66 credits, landed differently: 720p costs 10 credits per second (30 / 50 / 100 / 150 credits at 3 / 5 / 10 / 15 seconds), and 1080p costs 15 credits per second (75 / 150 at 5 / 10 seconds). These are report-grade readings from one account on one date, screenshot on file — not an officially published rate table, and not re-verified by a second party. The panel offered a maximum of 15 seconds and showed 4K as coming soon. The 9 and the 10 are both kept here unreconciled: the older reading may predate 4.0's pricing or may not describe 4.0 at all, and nothing on the record says which.

The arithmetic that follows is ours, not the page's. At 10 credits per second, the observed 66-credit package funds 6.6 seconds of 720p in total: one 5-second clip at 50 credits leaves 16, and 16 will not open the shortest clip the panel offered, 3 seconds at 30 credits. In 1080p it funds nothing, since a single 5-second clip is priced at 75. A 30-second 4.0 generation — a length the product page now advertises under native 30s-generation, though the 26 October panel did not offer it — would cost 300 credits by linear extrapolation from the per-second rate. That figure is extrapolated, never measured.

Two places where the vendor disagrees with itself

Both are official readings, dated, and neither is resolved here. First, resolution: the comparison table read on 4 November 2026 puts a dash against 1080p/4K in the free column, yet the logged-in generator of 26 October 2026 offered 1080p to that free-package account and priced it at 75 credits for 5 seconds. Second, the Kling 4.0 tier question: the comparison table has no row named Kling 4.0, and the Model Benefits of all four paid tiers listed only VIDEO 2.6 - Voice Control on 4 November 2026 — so which tiers can run 4.0 cannot be read off the public page on that date, even though a banner captured on 26 October had pointed at Pro or Higher. The banner itself moved between the two dates: Get 10% Off with Pro or Higher then, Yearly 34% Off and Monthly 12% Off now.

The three grades

Three grades apply to every row this site holds, and they are the reason a figure never travels here without one attached. vendor means the company stated it in its own material. report means someone using or hosting the product said it. single-source means exactly one person said it and nothing corroborates it. The field is not decoration: it decides how much weight a row can carry, and it is why the most technical-sounding claim in this dataset is also the one it will not let you lean on.

The most precise answer came from the weakest source

Here is the full chain of what the dataset holds on one narrow question — output resolution — arranged by grade rather than by date:

The grades run from strongest to weakest going down, and the answers get more specific going down. The vendor with the most authority over the question wrote nothing at all. The host that made the boldest claim contradicted itself one click away in its own pricing table. The one claim that names a bit depth — the most technical, most checkable-sounding line in the set — is the one row this dataset would not let you lean on.

That inversion is the reason "who said it" outranks "how much". A precise number is not evidence of a good source; on this question the two moved in opposite directions.

Vendor rows are not a single solid block either

It would be tidy if the top grade were internally consistent. It is not, and the dataset records both cases rather than smoothing them:

• One vendor contradicts itself. The 4K claim above is flagged as contradicted by the same platform's own plan table. A vendor statement is one grade; it is not automatically one coherent statement.
• Vendor and report disagree, unresolved. On API availability, the vendor's launch post said "coming soon" on 2026-07-31. A developer reported it had opened on 2026-07-16, in a row observed 2026-07-15. Both rows are kept, the conflict is flagged, and the dataset does not adjudicate — because at this grade of evidence, taking a side would be a guess dressed as a finding.

So the useful habit is not "trust official sources". It is "grade the source, then check whether it agrees with itself and with the other grades".

What to do with a Kling AI figure

Applied to a search for Kling AI's free access, three rules follow. None of them needs a number for Kling AI, and none of them invents one.

1. Attach the grade before the amount. When you find an allowance figure, write down next to it whether the vendor said it, a host or user reported it, or one person mentioned it with nobody else backing it. In this dataset that field decides how much weight the row can carry — it is why the 4K + 10-bit line stays a single row instead of becoming the headline.

2. "The official page doesn't mention it" is a finding, not a gap to fill in. The vendor row for resolution says not stated, and that row is dated 2026-07-31. It is a vendor row in the dataset. Silence from the authoritative source is information: it means the precise answer people quote for it is coming from somewhere weaker.

3. A single-source claim needs a second source before you plan around it. One row, one person, no corroboration — the dataset's own note says the source was under NDA. That is not the same as false. It is the same as unusable for anything you would be annoyed to get wrong.

Grading a number yourself

The shortest path to a Kling AI answer you can defend, in the order this dataset does it:

1. Find the figure, and record who published it — vendor material, a hosting platform, a forum post, a colleague.
2. Put it in one of the three grades above. If you cannot tell which grade it belongs to, treat it as report at best.
3. Add today's date. A row without a date is the one row type this dataset does not contain, because it would not be usable.
4. Look for the same claim from a different grade. If the vendor says one thing and users say another, keep both — do not average them and do not pick the one you like.
5. Check the unit before the amount, and whether the grant refills. In this dataset: 5 generations per account per day on Doubao (2026-08-11) sits in the same model's table as a one-off 20 credits on Mina Labs (2026-08-11). Same day, opposite behaviour over time, both filed without reconciling them.
6. Only then write the number down as your answer — with the grade and the date still attached.

Step 4 is the one most people skip, and it is the step that produced the contradiction rows above. Grok Imagine Free covers what to do once you have a graded number and want to know how long it stays true; Sora 2 Free covers the six fields to fill in for a first measurement of any model.

Do Kling free credits expire?

Yes — on the account observed here, the free trial package carries an expiry date of 3 November 2026. Both facts come from the same banner inside a logged-in session on klingai.com, read on 2026-10-03: the package holds 66 credits, and the same banner carries the date. Both are report rows — read by this site's owner, screenshot on file, not independently re-verified by a second party. What "expire" means here is worth spelling out, because the row's own note does it: the free grant is time-limited rather than a recurring daily allowance. There is no daily balance that drains and refills; the whole package has one date printed on it. What the dataset does not hold is anything about the other side of that date — whether unused credits disappear on 2026-11-03, whether the account is then offered a new package, or whether the date can be extended. No row addresses the aftermath, so none of it is claimed here. As of the 4 November 2026 check the printed date has passed; nothing here has observed that account since, so whether the credits lapsed with it remains unrecorded. For contrast within the same dataset, Runway's free deposit is stated by the vendor as a one-time 125 credits that does not expire (runwayml.com pricing page, 2026-10-02): the grant with an explicit expiry date and the grant with an explicit no-expiry statement sit on two different channels, and neither row says anything about the other.

Scope of what this page says

The Kling rows were read from klingai.com on 2026-10-03: the pricing URL from an unauthenticated headed browser, and the trial package, its expiry, the plan table and the 720P rate from a logged-in session on the same day, with screenshots on file and no independent re-verification by a second party. The additions dated 26 October 2026 come from a logged-in generator session, and those dated 4 November 2026 from the membership page read without any login — each figure keeps the date and the session it was read in. They describe one account on one channel on one date — and none of them names a Kling version, which is why a search for Kling 3.0 free, Kling 2.5 free or Kling free credits is answered by the same rows. Whether a specific version is reachable on the free grant is not something this dataset can show, and no figure here is offered as a per-version allowance.

The evidence-grading discussion above draws on the Seedance 2.5 rows, read across Jimeng, Doubao, Doubao Work, Mina Labs, Lovart, BytePlus ModelArk and vendor material between July and August 2026. None of those describes Kling AI, and nothing on this page is an industry-wide finding. The 68 / 56 / 1 split is a count of this site's own rows, not a measurement of how free-tier information is distributed in general.

The full dataset — every row this site holds, each carrying channel, observation date and source type — is on the data page in three formats, free to use under CC BY 4.0 with attribution. To turn an allowance and a usable rate into an attempt count, the attempts-needed calculator uses only figures from that dataset and estimates nothing on your behalf.

Not confirmed on this page: no Kling version is named in any row this site holds, so a version-qualified claim — Kling 3.0 free, Kling 2.5 free — cannot be confirmed here; the 66-credit package, its expiry and the 9-credits-per-second rate come from one account on one channel on 2026-10-03 and are not independently re-verified by a second party; when they were read on 2026-10-03 the plan table rendered only inside a logged-in session, but on 2026-11-04 the same table was found rendering with no login at all (see the update above), so the login-wall claim no longer holds — the trial-package figures themselves remain one account's readings, not re-verified by a second party. Also not confirmed: any initial credit grant for the free tier (the comparison table shows a dash where the paid columns show numbers); whether free-tier output carries a brand watermark (only its removal is listed as a paid benefit, with no sentence either way); which plan tiers can run Kling 4.0 (no comparison-table row names it and no Model Benefits list mentioned it on 2026-11-04); the actual credit price of a 30-second 4.0 generation (extrapolated above, never offered in the observed panel); and what happened to the observed account's 66 credits after the printed expiry of 3 November 2026.

Update cadence: figures on this page change only when the dataset gains a dated row that touches them; every figure keeps its own observation date — check it before quoting as current.