Deepsolv vs Pencil AI: Pricing and Credit Costs

The same 50 credits can pay for about 500 images at Pencil's lowest image rate of 0.1 credit each (50 ÷ 0.1), or four 15-second 1080p videos at its lowest video rate of 0.7 credits per second (50 ÷ 10.5, rounded down), per Pencil's credit rules. Those two outcomes sit on one $14 plan.

That spread is why a headline price tells a buyer little. A plan that looks cheap for images can be spent on a handful of clips, and neither number says whether the ads you make are the ones worth testing.

The sections below turn Pencil's published rates into a cost grid and a weekly credit worksheet, so you can check your own volume before you pay. They then set Pencil against Deepsolv, which publishes this page and is one of the two tools compared. We rank which Meta ads to test next, and Pencil generates the assets, so the real question is which of those jobs you are paying for.

• Pencil's plans are Core at $14 a month (50 credits), Growth at $55 a month (250 credits) and custom-priced Pro. The cost per finished ad depends on the model, because every AI interaction spends credits.
• On Pencil's published rates, a 15-second 1080p video costs between 10.5 and 97.5 credits. On Core's monthly rate of $0.28 a credit, that is $2.94 to $27.30.
• Size a Pencil plan from a weekly credit budget, not from the headline price, because every AI interaction spends credits.
• For Meta advertisers who need to know which ads to make next, we recommend Deepsolv. We deliver ranked weekly test plans and execution-ready creative briefs, so you decide before you spend on production.

As of November 2026, Pencil sells three plans: Core at $14 a month, Growth at $55 a month, and Pro at custom prices. Annual billing lowers Core to $11 a month and Growth to $44 a month, per Pencil's pricing page. Pro has unlimited generations on committed consumption pricing.

Committed consumption pricing means a Pro buyer agrees a usage commitment in advance. Pencil renamed "generations" to "credits" on 23 September 2026 with no change to balances, so the two words mean the same thing on its credit-counting help page.

Core and Growth cannot exceed the monthly allocation, and unused credits expire each month. Pro is charged overage if it passes its quota. Pencil describes Core as for solo entrepreneurs, Growth as for fast-growing teams and Pro as for enterprises, per its Pencil vs Pencil Pro article.

In most cases, one Pencil credit buys one interaction with an AI model, and the cost ranges from 0.1 credit for a cheap image to 6.5 credits per second of video. Pencil states: "In most cases, a credit is every time you interact with an AI model to generate something.", per its credit-counting help page.

The published rates, all from that help page:

• Images: 0.1 to 3.2 credits each, depending on the model.
• Video: 0.7 to 6.5 credits per second at 1080p.
• Audio: 0.1 to 2.1 credits per minute.
• A standard feed variation costs 1 credit. A feed variation is one finished version of an ad template built from one row of a product feed.
• A custom automatic score costs 1 credit per run.
• Auto Resize, which converts an ad into a new format, costs 1 credit per new format.

Dividing each plan's price by its allowance gives the price of a credit. Applying that to Pencil's lowest and highest published rates gives the table below. This is our calculation from Pencil's published prices and rates as of November 2026, and it assumes you use every credit. Your real cost per credit is higher if you use fewer.

Working: Core $14 ÷ 50 = $0.28 and $11 ÷ 50 = $0.22. Growth $55 ÷ 250 = $0.22 and $44 ÷ 250 = $0.176. A 15-second video is 15 × 0.7 = 10.5 credits at the low rate and 15 × 6.5 = 97.5 credits at the high rate.

Within one asset type, the model choice moves the cost a lot: images vary 32-fold (0.1 to 3.2 credits) and video about 9-fold (0.7 to 6.5 credits per second).

Core on its monthly price also runs out fast. 50 credits covers four 15-second videos at the lowest rate (50 ÷ 10.5 = 4.8, rounded down). At the highest rate it covers about 7.7 seconds of video (50 ÷ 6.5), less than one 15-second clip. Growth's 250 credits cover 23 such videos at the lowest rate and 2 at the highest. These counts assume no re-rolls (regenerating a weak result), which also spend credits.

A test week of 3 concepts, each with 4 feed variations and one resize, uses 15 credits, or about 65 credits a month, which is more than Core's 50. Work out your own week, then multiply by 4.3. That is Deepsolv's model, not a typical week.

Our model for one test week, using Pencil's published rates:

1. Feed variations: concepts × variants × 1 credit.
2. Resizes: one credit per new format, per concept resized. AI Auto-Resize is in beta. A standard feed variation costs 1 credit regardless of how many formats it renders in, so skip step 2 if you resize through feed variations.
3. Re-rolls: your own allowance for regenerating a weak result.
4. Media: seconds of video × credits per second, plus images × credits per image.
5. Monthly total: weekly total × 4.3.

Worked example, which is our assumption and not a Pencil figure: 3 concepts × 4 feed variations = 12 credits, plus 3 resizes (each concept resized once) = 15 credits a week. Then 15 × 4.3 = 64.5, about 65 credits a month. That is already over Core's 50. It is about 26% of Growth's 250 (64.5 ÷ 250), and it excludes re-rolls, images and video.

Check yourself against a simple test. Pass: your monthly total is at most 80% of the plan allowance. That is our recommendation, to leave headroom for re-rolls. Fail: it is above 80%, so step up a plan or lower the volume.

The two jobs can also sit together: we decide the test, and a production tool makes the assets. Our comparison with another AI ad generator is in Deepsolv vs AdCreative.ai.

On the Deepsolv side, we analyze the creative elements of competitor ads, such as hooks, messaging, offers, formats and visuals. What competitor ad volume actually tells you explains how to read that activity without over-reading it. See what a winning ad costs and who makes it for the production side.

Run these five checks, each with a pass or a fail, before you sign for either tool.

1. Can you state your monthly volume in credits? Pass: you have a number from the worksheet above. Fail: you only have a plan name.
2. Does the cheapest plan hold your busiest month? Pass: the total is within 80% of the allowance. Fail: you would need to buy the next plan mid-month.
3. Do you know the decision before you spend? Pass: every asset you generate traces to a written hypothesis. Fail: you are generating first and deciding later. A/B test hypothesis for Meta ads gives a template.
4. Do you have the cancellation and refund terms in writing? Pass: you can state the notice period and whether a mid-term upgrade is refundable, and you have them in writing before paying. Fail: you are relying on the plan name.
5. Is the price scoped to your real setup? Pass: the quote names your brands, team and workflow. Fail: the quote is a plan name only. For Deepsolv, our pricing explanation covers what a quote is scoped to.