How this guide is built: Opus Magnum is an open-ended optimisation game, not a “find the one answer” game. Every puzzle has infinitely many working solutions, and the moment yours runs the game plots it on a histogram against everyone else’s for cost, speed and size. A flat list of “here is the solution to puzzle 14” would therefore be both meaningless and a waste of the game — which is exactly why several of the community solution dumps this guide drew on have been removed from Steam, and why the survivors openly admit their answers “aren’t optimal.” So this guide teaches the systems instead: the atoms and glyphs, the arm-and-timeline programming model, the three scores and how they trade off, the golfing techniques that crack any puzzle, a worked spine of representative puzzles paraphrased as lessons, the Sigmar’s Garden minigame, and a spoiler-walled story summary. Learn the engine and you can build a machine we never mention. Facts are cross-checked against the official Steam page (app 558990), ten community guides and the official trailer.
What is Opus Magnum?
Opus Magnum is a $9.99 open-ended engineering puzzle game by Zachtronics — the studio behind SpaceChem, Infinifactory, TIS-100 and SHENZHEN I/O — released on 7 December 2017. You play Anataeus Vaya, the freshly hired Head Alchemist of House Van Tassen, and your job is to build transmutation engines: machines on a hexagonal grid that grab raw atoms, run them through alchemical glyphs, and assemble finished products — remedies, gemstones, weapons, threads and more.
What makes it special is that it is widely considered the friendliest Zachtronics game. There is unlimited space, unlimited parts, and no part-count limit; you cannot really get walled out of progress, because almost any machine that produces the product counts as a solve. The difficulty — and the addiction — lives entirely in optimisation. One community ranking put completing Opus Magnum at a gentle 4/10, then noted that chasing top histogram scores “elevates every puzzle 2–3 points.” That is the whole shape of the game: get it working, then make it beautiful.
It is also still alive. In March 2026 Zachtronics shipped a substantial DLC, De Re Metallica (17 puzzles across 3 chapters, 3 new glyphs and a new solitaire variant, a prequel about the alchemist Saverio Daas), alongside a Complete Edition. Everything in this guide applies to the base campaign and carries straight over to the new content.
Why this guide doesn’t just list the solutions
It is a fair question, so here is the honest reasoning. A puzzle-by-puzzle answer key for Opus Magnum is the wrong tool for three concrete reasons:
- There is no “the” answer. A machine that makes the product is the beginning, not the end. The game immediately asks you to make it cheaper, faster or smaller, and those goals fight each other. Our own sources prove it: one ranks every puzzle three ways — Cheap, Quick, Neat — and another labels its variants “Cheaper,” “Faster,” “Smaller.” A single solution would hide the only part that matters.
- The dumps don’t survive. Three of the ten community guides behind this article have been taken down by Steam for guideline violations, and two of the ones still up say outright that their solutions are “not optimal, just braining out.” Copying a screenshot of a mediocre machine is low value and short-lived.
- It defeats the game. The reward in Opus Magnum is the click of an engineering idea — realising you can pipeline the output, or that one arm can do the job of three. A GIF of someone else’s machine deletes exactly that moment.
The durable alternative is to learn the engine fluently. The next sections — the atoms, the glyphs, the arms and timeline, the three scores, and the golfing techniques — are built so that when you stall you can diagnose why and reach for the right idea, on any puzzle, in any version, including the new DLC.
The board: atoms and bonds
Every puzzle gives you one or more reagents (input atoms, dispensed from fixed tiles) and asks for one or more copies of a product (a specific arrangement of bonded atoms). Your machine has to turn one into the other. The atoms fall into a few families:
| Atom | What it is |
|---|---|
| Salt | The universal connector. Any of the four cardinal elements can be turned into salt, and two salt feed the Animismus glyph. Salt is also the wild card in Sigmar’s Garden. |
| Air, Water, Fire, Earth | The four cardinal elements. You can rotate between them around a salt hub — the famous “Van Berlo’s Wheel.” |
| Quicksilver | Mercury — the fuel that the Projection glyph burns to promote a metal one rung up the ladder. |
| Vitae & Mors | Life and death. The Animismus glyph turns two salt into one of each; many tonics and potions need them. |
| The six metals | Lead → Tin → Iron → Copper → Silver → Gold — a fixed ladder you climb with Projection or Purification. |
| Quintessence | The fifth element. Unification fuses one each of air, water, fire and earth into it; Dispersion splits it back. |
Atoms join with bonds. The product diagram in the top-left of the editor shows exactly which atoms must be bonded to which — read it like a blueprint, because reproducing that bond pattern is the puzzle.
The glyphs: the verbs of alchemy
If atoms are the nouns, glyphs are the verbs. A glyph is a symbol you place on the grid; when the right atom is dropped onto it, the reaction happens automatically. The campaign feeds them to you one at a time — here is the full set so you know what each is for:
| Glyph | Effect |
|---|---|
| Bonding / Unbonding | Bonding joins two adjacent atoms; Unbonding breaks a single bond. The two most-used glyphs in the game. |
| Triplex Bond | Bonds three Fire atoms into a triangle (fire only). Needed for several fiery products. |
| Calcification | Turns any cardinal element into Salt. Your way of making the universal connector on demand. |
| Duplication | Copies an element onto an adjacent Salt. The engine behind Van Berlo’s Wheel and any “make N copies” puzzle. |
| Projection | Consumes a Quicksilver to promote a metal one step up the ladder (lead→tin, tin→iron, …). |
| Purification | Fuses two of the same metal into one of the next metal up — no quicksilver needed, but it eats two inputs. |
| Animismus | Consumes two Salt and produces one Vitae + one Mors. Feed it at the right rate or it stalls. |
| Unification / Dispersion | Unification merges air+water+fire+earth into Quintessence; Dispersion reverses it. |
| Disposal | Destroys any atom dropped on it — a sink for the waste a reaction leaves behind. |
Arms, instructions and the timeline
You move atoms with mechanisms. There are only a few:
- Arm — a gripper on a pivot that rotates 60° at a time. The basic worker. Longer arms cost more.
- Double, Triple and Hex (six) arms — one pivot, several grippers spaced around it, so one rotation moves several atoms at once.
- Piston — extends and retracts along its own axis (this is translation — it moves an atom in a straight line without turning it).
- Track — rails you lay down so an arm can slide along them.
Each mechanism runs a little program — a row of instructions: grab, drop, rotate clockwise/counter-clockwise, extend/retract a piston, advance/retreat along a track, pivot, reset (snap back to the start state) and repeat (loop). Here is the single most important thing to understand:
Every arm loops its program forever, and they all share one clock. The machine’s period is the instruction count of the longest arm; shorter arms automatically pad with waits to stay in sync. Empty space at the start of an arm’s timeline is a free start-up delay; gaps in the middle or end cost you cycles every single loop.
That one fact is the key to speed. If you want a faster machine, find the arm with the longest program — the bottleneck — and shorten it. Polishing an arm that already finishes early does nothing.
A few editor tricks will save you hours: drag a fresh arm from the tray on top of an existing arm to change its type while keeping its program; hold Ctrl and drag to copy parts or instructions; box-select to rotate or copy a whole cluster at once; and right-drag to pan around. Undo (Ctrl+Z) and redo (Ctrl+Y) are unlimited — experiment freely.
The three scores — and why you can’t win all three
When your machine finishes, the readout shows three numbers, and the game ranks each on its own histogram against the whole community:
- Cost (in g) — the total price of the parts you used. Fewer and simpler mechanisms and glyphs = cheaper.
- Cycles — how long the machine takes to finish all the required products. Parallelism = faster.
- Area — how many hexes your machine occupies. Tighter packing = smaller.
The trap for newcomers is trying to win all three at once. You cannot, in general. The cheapest machine usually uses a single arm doing everything slowly; the fastest machine usually uses many arms and lots of parts, so it is expensive and large; the smallest needs awkward, often slow routing to stay compact. The right mindset is: solve the puzzle once, then pick one metric and rebuild for it. Many veterans keep three separate saves per puzzle — one cheap, one quick, one neat.
Golfing: the techniques that crack any puzzle
“Golfing” is the community word for minimising one of the three scores. These transferable moves — distilled from the optimisation guides this article draws on — are what you actually reach for when you’re stuck:
- Get any machine working first. A finished ugly solve teaches you the reaction sequence; only then is it worth optimising.
- For Cost, strip down. A reliable rule of thumb: almost every puzzle can be solved by one arm; if one arm can’t reach, add a piston; only if that fails, add a little track. Every extra arm and glyph is money.
- For Cycles, parallelise — but respect the input limit. A reagent tile dispenses at most one atom every two cycles, so the way to go faster is more arms, each doing one job (grab, process, reset) in a tight loop, rather than one arm doing everything.
- Attack the bottleneck. The longest arm sets the period. Shorten that arm’s program; ignore arms that already idle.
- Use multi-gripper arms to skip resets. A loop that grabs, rotates and drops then has to rotate back wastes cycles; a double, triple or six-arm can pick up the next piece on the way round instead of resetting.
- Choose translation vs rotation deliberately. A piston/track move is one cycle per hex and keeps the atom’s orientation; a rotation is one cycle per 60° and is better for moving something quickly across a longer arc.
- Check steady state. When the machine is humming, outputs should be produced at the same rate inputs are consumed. If a reagent tile sits idle, you have slack to go faster.
- For Area, refuse to waste hexes. Don’t let an arm sweep over empty space; tuck glyphs right against the work; reuse a single bonder for many bonds instead of placing several.
- When you’re truly stuck, walk away. As one guide put it, sleeping on a puzzle and coming back fresh “isn’t giving up — that’s the Zachtronics way.”
A tour of the campaign
The campaign is a teaching curriculum. You don’t need solutions for it — you need to know what each chapter is training so you can recognise the pattern:
- Prologue — The Transmutation Engine. The single puzzle Stabilized Water teaches the core loop: grab, bond, drop, repeat.
- Chapter I — Life in a Great House (House Van Tassen). Refined Gold, Face Powder, Waterproof Sealant, Hangover Cure, Airship Fuel, Precision Machine Oil and more. This is where Bonding, Calcification and Projection become second nature.
- Chapter II — Unprecedented Actions (House Colvan). Rocket Propellant, Explosive Phial, Armor Filament, Surrender Flare… bigger molecules and your first threads (repeating chains).
- Chapter III — Exile and Thievery (the Downriver Quarters). Separations and purifiers (pulling atoms apart), Climbing Rope Fiber, Very Dark Thread. Working backwards from a product becomes the skill.
- Chapter IV — An Uncommon Ally. Sword Alloy, Purified Gold, Alchemical Jewel, Golden Thread — metals and gemstones at scale.
- Chapter V — Opposing Plots. Timing Crystal, Voltaic Coil, Universal Solvent — Quintessence and the hardest campaign assemblies.
- Production (appendix). Money-making puzzles (Silver Paint, Viscous Sludge, Fragrant Powder…) with extra rules — a different kind of constraint puzzle.
A worked spine: one lesson per mechanic
Rather than dump solutions, here is the shape of a handful of representative puzzles — in plain words, so you keep the satisfaction of building them. Each one teaches a mechanic you’ll reuse forever.
- Stabilized Water (Prologue) — grab a water atom, bond a second one to it, drop the pair. The lesson is simply that a program loops: write it once and it runs forever.
- Refined Gold (Ch. I) — feed lead and quicksilver through a Projection glyph, over and over, to climb the ladder to gold. Lesson: one fuel atom is consumed per promotion, so plan your quicksilver supply.
- Face Powder (Ch. I) — Calcify cardinal atoms into salt and bond them into the required shape. Lesson: when a product wants salt you don’t have, make it.
- A thread (Very Dark Thread / Golden Thread) — have one arm bond a single new atom onto an ever-growing chain each loop while an output track hauls the chain away. Lesson: this pipeline pattern (build the screenshot above) is how you mass-produce anything long.
- An Animismus product (a tonic) — turn two salt into vitae + mors with Animismus, then bond them into the product. Lesson: a two-in/two-out glyph has to be fed at the right rhythm or it deadlocks.
- A Quintessence product (Ch. V, e.g. Universal Solvent) — assemble one each of air, water, fire and earth and feed Unification. Lesson: timing four different atoms to arrive together is the real puzzle.
- Van Berlo’s Wheel (Journal) — an arm plus a Duplication glyph spinning a salt hub to emit air→water→fire→earth in rotation. Lesson: this is the most reusable construction in the game — learn it once, deploy it everywhere cardinals are needed.
Sigmar’s Garden: the solitaire minigame
Tucked inside the story is Sigmar’s Garden, a hexagonal marble solitaire. You remove matching pairs that each have at least three free neighbouring spaces, until the board is clear. It feels like luck, but a little discipline wins most boards:
- Clear Lead first. Find the dark marble marked with the “h” (lead) symbol and pair it with one of the five Quicksilver marbles before doing anything else. Spend your quicksilver carelessly and you can strand the lead, which loses the board.
- Take the metals in order.
Lead → Tin → Iron → Copper → Silver → Gold. Each metal except gold needs a quicksilver pairing; gold is removed last, on its own. - Keep the cardinals and salt even. You start with eight of each cardinal and four salt. Same-colour cardinals pair together; salt is wild (it pairs any cardinal, or another salt). Always keep an even number of each or you’ll be left with one unmatched marble.
- Vitae pairs Mors, and in general remove blocking marbles before open ones, preferring any move that also frees the next metal in the sequence.
The Journal: the endless post-game
Finish the story and the real challenge opens up: The Journal, a stream of much harder puzzles published as Volume XCIX, Issues I–IX, five per issue — Van Berlo’s Wheel and Chain, Reactive Cinnabar, Ravari’s Wheel, the Animismus Buffer, the Quintessential Medium, and more. These are where the histogram becomes a genuine competition; even dedicated solvers leave a few marked “unsolved.” If you enjoyed the campaign’s optimisation, this is the main course — and it’s where the techniques above stop being optional and start being the only way through.
The story (light spoilers)
You can ignore this and lose nothing mechanical, so here is only the shape of it. Vaya begins as the celebrated new alchemist of House Van Tassen and is quickly drawn into a feud with rival House Colvan. Events turn against him — exile, poverty, and a stint surviving in the Downriver Quarters — before an uncommon ally and an elaborate counter-plot pull him back into the Houses’ intrigues, with a rival alchemist, Verrin Ravari, quietly steering more of it than anyone realises. The cast (Henley Servin, Concordia Lem, Captain Gelt, the thief Nils Whittle, Clara Soria) and the chapter titles — Life in a Great House, Unprecedented Actions, Exile and Thievery, An Uncommon Ally, Opposing Plots — trace the arc. The 2026 DLC De Re Metallica tells a separate prequel about the maverick researcher Saverio Daas. That’s as much as we’ll spoil; the dialogue is worth reading in place.
Watch it in motion
If you’re new and just want to see what a transmutation engine looks like running, the official trailer below gives the clearest two-minute picture of the loop — atoms in, glyphs firing, arms looping, product out.
The one habit that matters
Opus Magnum rewards a single habit above all: build it ugly, then make it good. Don’t stare at a blank grid trying to design the perfect machine — place an arm, grab an atom, get something producing the product, and let the histogram tell you where you stand. Then choose one number — cost, cycles or area — and rebuild for it using the techniques above. That loop, repeated, is the whole game, and it’s why a fixed answer key would have nothing to offer: the answer was never the point. The engine is.