Choosing between them
Four of these five hand you an empty grid
The most useful thing to know before printing a logic sheet is not how hard it is. It is how much of the grid arrives already filled in, because that decides whether the page looks like an invitation or a wall.
Why these four, and not every puzzle that needs thinking
A logic puzzle here means one where the answer follows from the clues by deduction and can be checked against a single correct grid. That is why Sudoku, KenKen, Kakuro, and Futoshiki are on this page and a word search is not: a word search is a search, and finding a word is not a deduction. Cryptograms sit under word puzzles for the same reason, even though solving one is genuine reasoning. The dividing line is whether the grid can be proved right, because that is what makes a printed answer key worth anything.
| Puzzle | Sheet measured | Squares you fill in | Numbers printed | Other clues printed | Filled before you start |
|---|---|---|---|---|---|
| Sudoku | 9x9, Medium | 81 | 32.0 | None. The grid is the whole clue. | 40% |
| KenKen | 5x5, Medium | 25 | 2.4 as one-cell cages | 11.6 cage totals | 10% |
| Kakuro | 9x9, Medium | 39.5 white of 81 | 0 | 21.8 clue squares, 25.3 run totals | 0% |
| Killer Sudoku | 9x9, Medium | 81 | 0 | 33.7 cage totals | 0% |
| Futoshiki | 5x5, Medium | 25 | 0.5 | 9.8 inequality signs | 2% |
Sudoku, KenKen, and Futoshiki figures are the mean of 8 sheets generated at build time by the same engines the builders use. Kakuro is not generated in the browser, so its row is the mean of all 110 nine-by-nine medium puzzles in the shipped bank. The last column is numbers printed divided by squares to fill.
Sudoku is the outlier, and it is the reason Sudoku is the one everybody starts with. It is the only puzzle of them that begins by giving you answers: two fifths of a Medium grid is already correct before you pick up a pencil, and your job is to extend a pattern that is visibly there. The other three tell you nothing about any individual square. They describe relationships, and the numbers are what you get for working the relationships out.
KenKen looks like an exception because a few numbers do appear, but they are given numbers in disguise: a one-cell cage is a square with its own answer written above it. A Medium 5x5 averages 2.4 of them out of 11.6 cages, and they are the cheapest way a generator has of making a grid solvable. Futoshiki sits at the other extreme with 0.5 numbers on a Medium 5x5, which usually means none at all. That looks like a printing fault the first time you see it and it is not: an empty grid with the right signs on it still has exactly one answer. Easy keeps 4.1 numbers on purpose, because a beginner needs somewhere obvious to start.
The same four sheets, drawn at the same size
Real sheets, generated at build time by the live engines. Nothing here is a mock-up of a grid.
Difficulty means a different thing on each of these pages
Four builders, four difficulty controls, and no shared meaning between them. A Hard Kakuro and a Hard Futoshiki are not graded on the same scale and never were, so the honest thing is to say what each control actually moves.
| Builder | What the difficulty setting changes | What it does not promise |
|---|---|---|
| Printable Sudoku | The number of clues left on the grid. Easy 40, Medium 32, Hard 28, Expert 24, hit exactly every time. | It does not promise which deductions a sheet needs. Clue count is a proxy for difficulty, and it is the loosest definition on this site. |
| Printable KenKen | The shape of the cages: how large they get, how many one-cell freebies are allowed, how bent the cages are, and which operators appear. | It does not promise a named solving technique either, though cage size is a closer proxy than clue count because a four-cell cage genuinely has more combinations to sift. |
| Printable Kakuro | Which shelf of the bank the sheet is drawn from. Each puzzle was graded and solver-checked when it was made, and the grade travels with the record. | It is fixed at the point the bank was built, so it cannot be re-tuned per sheet. Grid size is a separate control and a large easy Kakuro is a long puzzle, not a hard one. |
| Printable Killer Sudoku | The hardest deduction the sheet forces you to find, from a ladder of five: cage combinations, the rule of 45 on one house, naked and hidden subsets, innies and outies across a band, and following a candidate to a contradiction. | It does not promise how many cages you will have to read. Harder levels have fewer and larger cages, which is less to look at and more to work out. |
| Printable Futoshiki | The hardest deduction the sheet forces you to find, from a ladder of five: scanning a line, reading one sign against a filled cell, following bounds along a run of signs, naked and hidden subsets, and following a candidate to a contradiction. | It does not promise how long a sheet takes. An Expert 5x5 that needs one contradiction test can fall faster than a Medium 7x7 that needs thirty patient rounds. |
Ranked by how much the label tells you, Futoshiki is first and Sudoku is last. Futoshiki names the hardest deduction the sheet forces and regenerates until the puzzle genuinely needs it. Sudoku counts clues, and clue count is only loosely related to how a grid behaves: two Medium sheets with the same 32 numbers can want quite different work. That is a real weakness of the Sudoku page rather than a quirk to defend, and it is on the list to fix. What Sudoku's control does give you is predictability, which has its own value: 32 clues is exactly 32 clues, every time, which is what a teacher printing thirty identical-looking sheets actually wants.
Two of them need arithmetic. Two need none at all
This is the question that decides most classroom and family choices, and it splits them cleanly down the middle.
- No arithmetic: Sudoku and Futoshiki. The digits are distinct symbols and nothing is ever added to anything. You could print either with animals in place of the numbers and the puzzle would be unchanged, which is exactly what a Picture Sudoku is.
- Real arithmetic: KenKen and Kakuro. KenKen asks for the four operations inside a cage, and the grid size changes the sum: a two-square cage marked 12x has one answer on a 5x5, where only 3 and 4 reach it, and two answers to weigh on a 6x6, where 2 and 6 do as well. Kakuro asks for addition plus a feel for combinations: a three-square run totalling 6 can only be 1, 2, and 3 in some order, and that single fact opens more Kakuro grids than any other.
What that means when you are printing for someone else
- A child who is still shaky on adding will stall on Kakuro for reasons that have nothing to do with logic. Give them Sudoku or a small Futoshiki and the deduction is the only obstacle left.
- KenKen is the one to reach for when you want arithmetic practice that does not look like a worksheet of sums. The maths is real and it is in service of the puzzle.
- Kakuro is the most demanding of them to start and the most rewarding to keep. It is a poor first sheet and an excellent tenth.
- Futoshiki teaches one specific thing better than any other sheet here: that the greater-than sign is information, not notation.
One answer and no guessing are two different promises
Most printable puzzle sites, when they say anything at all, say their puzzles have a unique solution. That is worth having and it is not the promise a person solving in pen cares about. A grid can have exactly one answer that is only reachable by writing a number in, following it for ten minutes, and finding it breaks. On screen that is an undo. On paper it is a ruined sheet.
| Puzzle | Exactly one answer? | Reachable without guessing? | The catch |
|---|---|---|---|
| Sudoku | Yes. Every generated puzzle is counted before it is shown, and one with a second answer is thrown away. | Not proved. | Clues are removed while the answer stays unique, which is not the same as staying findable. |
| KenKen | Yes. Cages are strengthened until the solution count is exactly one. | Not proved. | In practice 4x4 and 5x5 sheets are comfortable; a Hard 6x6 can ask for a long combination sift. |
| Kakuro | Yes. Every one of the 110 nine-by-nine medium records in the bank carries a passed solver gate with exactly one answer. | Not proved. | The bank records uniqueness, not the route to it, so a hard sheet can still need a leap. |
| Killer Sudoku | Yes, and like Futoshiki it falls out of the solve rather than being a separate check. | Guaranteed. | Cages are split until a solver restricted to the level’s own techniques can finish, then merged back while it still can, so every cage boundary that survives is one the solve needs. |
| Futoshiki | Yes, and it is a consequence rather than a separate check. | Guaranteed. | The generator keeps a sheet only if a solver limited to the level’s own techniques finishes it, and a solve that never discards a possible value proves the answer unique on the way past. |
Two of the five make the second promise, and both make it the same way. A generator that has to run a deduction solver anyway, in order to decide what level a sheet is, gets uniqueness for nothing: because that solver never discards a value appearing in some answer, a solve that fills the grid has also proved the answer unique on the way past. The other three prove uniqueness by counting solutions, which is a completely different question and cannot tell you anything about the route.
This section used to say only Futoshiki managed it, and that closing the gap was the most valuable unfinished work on this part of the site. Killer Sudoku closed half of it: it was built on the same principle from the start, and it is the reason the column above now has two entries rather than one. Sudoku, KenKen and Kakuro still count solutions, and bringing them across remains the work worth doing next.
Which one to print, by the reason you are printing it
| If you are | Settings | Why that one | |
|---|---|---|---|
| A first logic sheet for a seven-year-old | Printable KenKen | 4x4, Easy, + only operators, 2 puzzles | Sixteen squares, small sums, and a finished grid inside ten minutes. |
| A sheet with no arithmetic in it at all | Printable Sudoku or Futoshiki | Sudoku Easy, or Futoshiki 4x4 Easy | Neither puzzle adds anything. Swap the digits for shapes and both still work. |
| Times tables practice that does not look like practice | Printable KenKen | 5x5, Medium, mixed operators, name and date on | Every multiplication cage is a factoring problem in disguise, and one sheet asks for a dozen of them. |
| A maths lesson on the greater-than sign | Printable Futoshiki | 5x5, Easy, 4 puzzles, 2 per page, rules page on | The sign stops being notation to memorise and becomes the thing that solves the problem. |
| Large print for a care home or an activity group | Printable Sudoku | Easy, large print, 1 per page | The most familiar of them all, so nobody needs the rules explained before they can start. |
| A long evening solve for an experienced solver | Printable Kakuro | 11x11, Hard, 1 per page, answers on | The one that most reliably runs past an hour on a single sheet. |
| A puzzle club sheet that is provably fair | Printable Futoshiki | 7x7, Expert, 1 per page | Hard enough to need a contradiction test, and still finishable without a single guess. |
| A cover lesson pack that fills a period | Printable Sudoku | Medium, 6 puzzles, 2 per page, rules page on, answers on | Three sheets of work, the rules on the front, and a key the cover teacher can mark from. |
What is missing from this hub, and why
Four builders is not a complete logic section, and the gaps are specific rather than a general promise of more to come. The largest is the clue-list deduction puzzle, the one with five suspects and a grid of crosses, which is what a lot of people mean when they type "printable logic puzzles" into a search box. It is missing because it is hard to do honestly: it needs a scenario generator, a clue generator, and a solver that proves the answer is reachable by pure deduction, and the version of that page which is quick to build is the version that prints unsolvable sheets.
Printable Nonograms
Picture logic puzzles driven by row and column number clues, where the finished picture is the answer key.
Not built yetPrintable Hitori
Shade the duplicates until no number repeats in a row or column, without isolating any square you keep.
Not built yetPrintable Slitherlink
Draw a single closed loop around numbered clues. Prints beautifully; proving a unique loop is the hard part.
Not built yetPrintable Binary Puzzles
Only noughts and ones, with three short rules. The natural entry point for anyone who finds Sudoku heavy going.
Not built yetPrintable Bridges (Hashi)
Connect numbered islands with the exact number of bridges each one asks for, without crossing any of them.
Not built yetThese are plans, not pages. None of them exists yet, so none of them is a link, and this list will shrink as they are built rather than growing to look impressive.
How these pages are made and checked
Print Puzzles is part of the Hamilton Digital Media Limited puzzle network, and these pages are written and reviewed by Brian Hamilton and Karan Hamilton. Every figure on this page is measured at build time from the same engines and the same puzzle bank the builders use, so none of it can drift away from what the tools produce. Each engine was checked with a temporary harness before its page went live, and what those harnesses assert is set out on how we make puzzle packs. If a printed sheet or an answer key does not behave, the contact page is the fastest way to tell us.
There is no account, no email step, and no software to install. Sudoku, KenKen, and Futoshiki puzzles are generated in your own browser; Kakuro sheets are chosen from a bank of pre-checked puzzles this site sends you. Either way the PDF is written on your machine rather than on a server, which is why a pack appears in a second or two and why nothing you type into a worksheet title ever leaves the page. The terms of use give broad permission to print, copy, and hand these out for family, classroom, tutoring, library, club, care home, and community use. What they do not cover is bulk republication or selling the sheets as a product of your own.