Devin vs Augment: Quota or Pooled Dollars

Devin and Augment Code both sell agentic coding, and they disagree about what a buyer should be charged for. One sells seats with a quota. The other sells a pool of dollars.

That difference changes which team each one suits, far more than any feature list on either page does.

Both pages were read with a headless browser on October 27, 2026 and checked a second time the same day. Totals marked as arithmetic are ours.

Devin sells a seat, then refreshes a quota

One note first. The address windsurf.com/pricing now serves the Devin plans page from Cognition, so buyers looking for the old Windsurf price list land here.

Individual plans are Free at $0, Pro at $20 a month and Max at $200 a month. The comparison table lists 1 member for each of those three.

Team pricing is split in two parts: $80 a month for the team plan plus $40 a month per full dev seat, with up to 200 users.

By our arithmetic, a ten developer team on that structure is $480 a month, or $5,760 a year. Enterprise is listed as Let us talk, with no published price.

The allowance is described rather than counted. Each paid plan comes with a usage allowance that refreshes automatically on a daily and weekly basis.

• Free: light quota to code with agents, limited model availability.
• Free: unlimited inline edits and unlimited Tab completions.
• Pro at $20: increased quotas, full model availability, free use of SWE-2, access to Devin Cloud.
• Max at $200: significantly higher quotas, no figure given.
• Overage on paid plans: extra usage purchased and consumed at API pricing.

Cost per message is said to vary by model, task size, complexity and reasoning required. So the quota is real but unquantified, as with Cursor's multiples of Plus.

Augment sells one price for the whole team

Augment prints two flat prices. Standard is $20 a month and Business is $100 a month, both marked flat with no per-seat charge.

Each covers up to 50 seats, and each includes usage equal to its price: $20 of usage a month on Standard and $100 on Business.

By our arithmetic, Standard at a full 50 seats works out at $0.40 per seat a month. Enterprise is custom user pricing with unlimited users.

A seat is defined as any team member with access to the workspace, and the included usage plus any top-ups is pooled across the whole team.

Pooling rather than per-seat allocation is the structural choice described in what Notion AI adds to a seat.

The usage unit is a dollar, with a stated markup

Augment is unusually explicit about what a dollar of usage pays for. Model inference is billed at the provider's public API list price.

On top of that sits a flat 40 percent service fee on LLM usage, with no fee on compute. Compute itself is Cosmos compute time.

So by our arithmetic, $20 of included usage spent entirely on inference buys about $14.29 of list-price model calls, since $20 divided by 1.4 is $14.29.

That is a markup disclosed in writing, which almost no AI vendor does. It also makes the plan directly comparable with running the same models yourself.

Compare it with the two pages that publish a unit price instead: Copilot at a cent per credit and Kiro at two cents.

Overage and expiry differ sharply

Augment allows top-ups as you go with no plan change, and states that top-ups not part of a paid plan are valid 12 months after the date of purchase.

Devin's answer is to purchase extra usage that is consumed at API pricing, which is a pass-through rate rather than a pack with an expiry date.

Neither page we read states whether unused monthly allowance carries forward. Several vendors do answer that, and the answers vary: see do unused AI credits roll over.

Augment also states that all paid plans exclude AI training on customer data, and that it undergoes annual SOC 2 audits.

Which structure suits which team

Augment is the cheaper sticker price for a group of light or occasional users, because a dormant colleague costs nothing. Fifty seats on a $20 plan is the extreme case.

Devin is the clearer fit where a few developers each run heavy agent sessions, since a per-seat quota that refreshes daily is harder to exhaust for one person.

The risk flips with the model. Pooled dollars can be drained by one enthusiastic user in a week; a per-seat quota wastes the capacity of anyone who logs in rarely.

Either way, convert the plan to an annual figure for your real headcount before comparing. That reading method is set out in how to read an AI credit pricing page.

Frequently asked questions

How much is Devin for a ten person team?

By our arithmetic, $480 a month. The page lists $80 a month for the team plan plus $40 a month per full dev seat, with up to 200 users supported.

How much usage does Devin Pro include?

No figure is published. Pro at $20 a month is described as increased quotas that refresh daily and weekly, with cost per message varying by model and task complexity.

Does Augment charge per seat?

No. Standard at $20 a month and Business at $100 a month are flat prices for up to 50 seats, with included usage and top-ups pooled across the whole team.

What is Augment's 40 percent service fee?

A flat fee on LLM usage, stated on the pricing page. Inference is billed at the provider's public API list price, then the fee is added. Compute carries no fee.

Which is cheaper for five light users?

Augment Standard, on sticker price: $20 a month flat against $280 for Devin at $80 plus five seats at $40. Heavy agent use can reverse that through top-ups.

Is Windsurf still sold separately?

The address windsurf.com/pricing served the Devin plans page from Cognition when we read it on October 27, 2026. We found no separate Windsurf price list there.

Bottom line

Augment publishes the most legible usage model of the two, because its unit is a dollar and its markup is written down.

Devin publishes the clearer team arithmetic, because $80 plus $40 a seat can be multiplied today. What it withholds is the size of the quota you are buying.

All figures read October 27, 2026. Related: what annual billing really saves and Otter against Fathom on free limits.