Claude Certification Blog

Pacing on the Claude exams: the published budgets spend all 120 minutes

All four guides publish a minute-by-minute pacing table, and every one of them spends the entire 120 minutes. No published plan has any review time in it.

All four sum to 1200 min reviewRun at 92%

7 min read

All four Claude exams run 120 minutes, and all four guides publish a suggested per-domain budget. Add any of them up and the total is exactly 120. Every published plan spends every minute you have, so none of them contains anything for going back to a flagged question. If you want review time, you have to create it by running each block under budget.

4of 4 publish a budget
120minutes each allocates
0left for review
92%to buy ten back

The budgets add up to all of it

CCDV-F suggests 18, 40, 4, 3, 20, 13, 10 and 12 minutes across its eight domains. CCAR-P suggests 20, 16, 23, 19, 17, 17 and 8 across its seven. CCAO-F gives 16, 26, 14, 20, 14, 18 and 12, and CCAR-F 32, 22, 24, 24 and 18. Four papers, four tables, and every sum is 120. That is a defensible way to allocate a fixed sitting, and it is also a plan that only works if you beat it — which is worth knowing before you are forty minutes in and wondering whether you are ahead.

All four publish one

ExamItemsPer itemPublished budgetWhere it is printed
CCDV-F532.3 min18/40/4/3/20/13/10/12Exam snapshot
CCAR-P631.9 min20/16/23/19/17/17/8Exam snapshot
CCAO-F602.0 min16/26/14/20/14/18/12Snapshot, inside the blueprint table
CCAR-F602.0 min32/22/24/24/18Exam-day checklist, not the snapshot

Two of them take finding: CCAR-F prints its split in the exam-day checklist rather than the snapshot, and CCAO-F’s is not a timing section at all but a column inside the blueprint table, headed “Time at 2 min/item”. Same fact, three different filing decisions — which is worth knowing before you conclude your exam does not have one.

The professional paper is the tightest: 63 items in the same two hours is about 1.9 minutes each, and it is also the most scenario-heavy of the four, so its stems take longer to read than its per-item average allows. That combination — most items, longest stems, same clock — is the real reason CCAR-P feels harder than its content alone would suggest, as CCAR-P after CCAR-F covers.

Buying back ten minutes

The correction is a single multiplier. Run each domain at about 92% of its published figure and the whole plan lands around 110 minutes, leaving nine or ten to spend where you need them:

The published CCDV-F budget, and the same blocks run at 92%PUBLISHED MINUTES, AND AT 92%D2 Applications37D5 Models18D1 Agents17D6 Prompt/ctx12D8 Tools/MCPs11D7 Security9D3 Claude Code4D4 Eval/debug3
The darker bar is the published figure; the lighter one is the same block at 92%, which is the number shown on the right. Two domains are already too small to trim.

Note that the trim does not fall evenly, and it should not. The 40-minute Applications and Integration block gives back three minutes on its own, because it is the block with room in it — it is 17 of the 53 items and the domain that decides CCDV-F. The four-minute and three-minute blocks give back nothing, because a domain worth one or two items cannot be run faster than it already is.

What the reserve is for

Not for a general re-read. Ten minutes across a 53-item paper is about eleven seconds an item, which buys you nothing if you spend it evenly. It is for the specific questions you flagged on the way past, and the value is entirely in having flagged them accurately.

Which makes flagging a skill worth practising rather than a habit worth having. The useful flag is the one where you narrowed the options and could not separate two of them; the useless one is the item you found hard and would find equally hard again with five more minutes. That distinction is the same one that decides items in the first place — see how to read a Claude exam question and multiple-response questions, where the select-count makes a returned-to item genuinely recoverable.

We had this in our own guides too

Both the CCDV-F and CCAR-P snapshots told candidates to bank any time they saved, from a table that left nothing to bank. Both now state the sum and carry the 92% correction, so the pacing advice and the pacing table finally agree with each other.

The same cut on the other three

The 92% rule is not specific to the developer paper. Applied to each published table it gives 15/24/13/18/13/17/11 on CCAO-F, 29/20/22/22/17 on CCAR-F and 18/15/21/17/16/16/7 on CCAR-P — landing at 111, 110 and 110 minutes, so nine or ten held back in every case.

Two of the four are easy to miss, and it is worth saying where they are. CCAR-F’s budget is in the exam-day checklist rather than the exam snapshot. CCAO-F’s is not a timing section at all: it is the last column of the blueprint table, headed “Time at 2 min/item”. Neither looks like the standalone timing tables the other two carry, which is exactly how a reader concludes their exam publishes no budget and starts building one.

Then time one full mock against the clock before you book. A budget you have never run is a hypothesis, and the thing it is most likely to be wrong about is how long the longest stems take you specifically — which is exactly what how many mock tests before the exam is for. The wider sequencing sits in our Claude certification study guide.

Key takeaways

  • Every published budget sums to exactly 120. All four allocate the whole sitting and leave nothing for review.
  • Run each block at about 92%. That returns nine minutes on CCDV-F and ten on CCAR-P.
  • The trim is uneven on purpose. Big blocks give minutes back; a three-minute block has none to give.
  • CCAR-P is the tightest paper. 63 items in 120 minutes, and the longest stems of the four.
  • All four publish one, in three different places. Two in the snapshot’s timing table, one as a blueprint column, one in the exam-day checklist.
  • A reserve is only worth what your flags are worth. Flag the items you narrowed, not the ones you found hard.

A budget you have never run is a hypothesis

Pacing is the one part of exam technique that practice papers measure directly, and the only way to find out whether 92% is comfortable or brutal for you is to sit one against the clock. Our mock exams use the same published item allocation as the real paper.

See the CCDV-F blueprint

Questions

Frequently asked

The follow-up questions people search next.

How long do you get per question on a Claude exam?

All four exams run 120 minutes. That is about 2.3 minutes per item on CCDV-F (53 items), 2.0 on CCAO-F and CCAR-F (60 each) and 1.9 on CCAR-P (63) — before you allow anything for going back over flagged questions.

Do the published time budgets leave room to review?

No, on any of the four. Every published table sums to exactly 120 minutes, so all of them allocate the entire sitting. Any minute you spend revisiting a flagged item has to be a minute you did not spend on a domain.

How do I make time to review flagged questions?

Run each domain block at roughly 92% of its published figure. On CCDV-F that turns the 40-minute Applications block into 37 and the 20-minute Model Selection block into 18, and hands back about nine minutes; the same cut on CCAR-P returns about ten.

Which Claude exams publish a per-domain time budget?

All four, though not all in the same place. CCDV-F, CCAR-P and CCAO-F carry theirs in the exam snapshot — CCAO-F’s as a column inside the blueprint table headed “Time at 2 min/item” — while CCAR-F publishes its 32/22/24/24/18 split in the exam-day checklist instead. Each one totals 120 minutes.

Is the tightest exam the one with the most questions?

Yes — CCAR-P puts 63 items into the same 120 minutes, which is about 1.9 minutes each and the least time per item of the four. It is also the most scenario-heavy paper, so the reading load per item runs the other way.

Keep reading

Related posts

Not affiliated with, or endorsed by, Anthropic or Pearson VUE. Details are summarised from publicly published program information and can change — always confirm against the official exam guide before booking.

We use cookies and privacy-friendly analytics to understand usage and improve Cred Farmer. Essential features work either way. See our Cookie Policy.