How to build The Ur-Sphinx

Nine mana buys a 10/10 that mills the whole table and casts what it finds, and the half of this card you will lean on costs nothing at all.

Nothing in this tribe costs less than three mana

Eighty-five Sphinxes can be cast inside a white, blue and black identity. Two of them cost three and there is nothing below that. Eleven cost four, twenty-three cost five, and twenty-nine cost six, which makes six the single most common price in the tribe by a wide margin. Fifteen cost seven, four cost eight, and the last one is the commander at nine.

Put that curve beside the eminence line and the card explains itself. A flat off would be a rounding error on a tribe of two-drops. On a tribe whose middle is a six-drop it moves the most common Sphinx in the format down to five, which is a full turn earlier in a deck that is already spending its first three turns on rocks. It turns Vexing Sphinx into a two-mana play and leaves Raffine, Scheming Seer at three, because a reduction that only eats generic mana cannot touch a cost that has none.

That discount is the reason the list can carry twenty-five of these things and still function. Take it away and the same ninety-nine cards are a pile of six-drops nobody would sleeve.

The tax break is free and the body costs nine, which are two different decks

Eminence is written to work while the card is in the command zone or on the battlefield, so nothing about the discount asks you to cast him. The attack trigger carries no such clause, and an ability with no zone clause works on the battlefield only. So the mill, and the free spells that come with it, cost the full .

Three cards in the list shorten that bill, and they work because of the shape of his cost rather than anything clever in their own text. is six generic mana next to three coloured, and generic is the only kind those reductions eat. Herald's Horn takes off , Urza's Incubator takes off , and Unesh, Criosphinx Sovereign takes off more. With all three out he costs , which is four mana for a 10/10 flier with the trigger attached.

Read the word other in the eminence line before assuming it stacks with itself. He never discounts his own casting cost, which makes him the one reduction in the deck that does not apply to the card printing it. The same three cards plus eminence do stack on everything else, and they stack all the way down: Consecrated Sphinx at a printed has four generic to give away and four reductions waiting, so it lands for .

The attack trigger reads every library at the table, and one of them is yours

Whenever one or more Sphinxes you control attack, each player mills that many cards, and for each player you may cast one of the cards they milled without paying its mana cost. Three attackers in a four-player game mills three off each of three opponents and three off you, and offers up to four free spells out of four separate piles.

Three things are worth being exact about. The cards are cast rather than put onto the battlefield, so a creature you take this way stays yours permanently, an X spell arrives with X set to zero, and an additional cost still has to be paid. Lands cannot be cast at all, so a milled land is a card you looked at. And the offer is one card per player rather than one card in total, which means the number of opponents is worth as much to this trigger as the number of attackers.

What it is not is dependable. You are reading the top few cards of somebody else's deck. A hundred-card list is usually around a third lands, so three milled cards contain something castable roughly nineteen times in twenty, and castable is a long way from worth casting. The trigger pays, and it pays in whatever the table happened to be sitting on.

Filling three graveyards is a real cost, and this list has no answer to the deck that wants it

The trigger is symmetrical in the one way that matters. Every attack puts cards into three graveyards you do not control, and across from a reanimator deck, or anything counting card types in a yard, you are their ramp. That is not a reason to leave this commander at home. It is a reason to know which pod you are in before the first attack.

The list does at least bill for its own half. Scholar of the Lost Trove casts an instant, sorcery or artifact out of your graveyard on arrival, Uldaros Theorix copies one nonland card of each type out of it, and Animate Dead and Reanimate both read any graveyard rather than only yours, so the two cards that answer a bomb somebody milled are the same two that steal one.

There is no graveyard hate in the ninety-nine and that is a gap rather than a subtlety. Against a table that wants the mill, Bojuka Bog and Soul-Guide Lantern come in over Arcane Sanctum and Gilded Lotus, which costs the mana base one tapped tri-land and the ramp its slowest rock.

Eight cards name a creature type, and one artifact makes the choice stop mattering

Cavern of Souls, Secluded Courtyard, Urza's Incubator, Herald's Horn, Kindred Discovery, Vanquisher's Banner, Icon of Ancestry and Kindred Dominance all ask you to name a type, and on all eight the answer is Sphinx. Maskwood Nexus is in the list to make that answer stop costing you anything. It reads that creatures you control are every creature type, so once it resolves everything you control is a Sphinx regardless of what its own type line says.

That pays in three places. Kindred Dominance destroys every creature that is not of the chosen type, which is already lopsided here and becomes entirely one-sided with the Nexus out, because a creature you dragged out of somebody else's graveyard is a Sphinx too. The attack trigger counts Sphinxes you control attacking, so that same creature adds to the mill number instead of watching. And Cavern of Souls fixes its type as it enters, which normally pins it to Sphinx spells only, while the Nexus quietly makes every creature spell you cast a Sphinx spell.

Thirty-six lands, and turn three is where the colours hurt

Eighteen cards in the ninety-nine want , and the cheapest of them can be cast on turn two, so the mana check measures this list against a turn-three deadline. Clearing that nine times in ten asks for thirty-six blue sources. The list is short of thirty-six and comes in at 83.9 per cent, which is the actual price of building this tribe in three colours instead of in mono blue. Black and white sit right on their targets rather than above them, because there was nothing left to pad them with.

The part worth reading twice is what the budget rung does to that number. Trading four fetchlands, City of Brass and Reflecting Pool for tapped duals, a filter land and two battle lands takes the same check to 90.6 per cent. The expensive mana base is buying speed, selection and an untapped first land. It is not buying colours, and in this list it buys fewer of them.

Path of Ancestry keeps a slot despite entering tapped every single time, because the mana it makes scries when it pays for a Sphinx, and a scry is not a small thing in this deck. Denzilore Fatehold puts a +1/+1 counter on every creature you control on any scry or surveil, and Meticulous Archive, Undercity Sewers, Sphinx of Foresight on each upkeep and Prognostic Sphinx on each attack are all triggers waiting for it.

Five cards that read as automatic here and are not

Eligeth, Crossroads Augur is a blue Sphinx that turns every scry into a draw, which sounds strictly better and is not. It is a replacement effect, so you never scry at all, and a trigger waiting for a scry never fires. Put it in and Denzilore Fatehold stops growing the team, Path of Ancestry goes back to being a tapped land, and Sphinx of Foresight and Prognostic Sphinx both lose half their text.

Fierce Guardianship and Flawless Maneuver are the free spells every Esper list reaches for, and both open with the words if you control a commander. This one lives in the command zone on purpose, so on most turns they are a three-mana counterspell and a three-mana indestructible trick. That is why the bracket 4 rung below skips the free counterspell that every deck at that power level runs.

Morophon, the Boundless is the cost reducer people name first for a nine-drop, and it takes the wrong three mana off. It reduces coloured mana only, so on it removes and leaves behind. Seven mana and then six is thirteen across two turns, against three-mana artifacts that eat the half of the cost that is actually expensive.

Medomai the Ageless takes an extra turn on combat damage and cannot attack during extra turns, so the turn it buys never produces a second mill trigger. Coat of Arms looks purpose-built for a list with twenty-five Sphinxes, and the board here is four or five large fliers rather than twenty tokens, and it hands the same bonus to every other tribal deck at the table.

The core list

Twenty-five Sphinxes, thirty-six lands and a nine-drop sitting on top, assembled at bracket 3. Hovering a name brings up the card, and each price shown belongs to the cheapest paper copy of it.

Load into the deck builder Copy decklist

“Load into the deck builder” opens this list in your account, ready to edit or save. “Copy decklist” gives you plain text for Moxfield, Archidekt or Arena — no account needed.

Going down to bracket 2 is mostly a mana base decision

Eighteen swaps take this list from at least $664.80 to at least $181.36, and eleven of the eighteen are lands. Those eleven are more than half the saving on their own. Cavern of Souls, four fetchlands, City of Brass, Reflecting Pool, Sea of Clouds, Morphic Pool, Otawara, Soaring City and Takenuma, Abandoned Mire come out for tapped duals, a filter land and two battle lands, and what you get back is the better colour count from two sections above. What you give up is tempo, because nine of the rung's thirty-six lands now enter tapped, and a deck aiming at a nine-mana turn feels a tapped land more than most.

Only seven cards change for a reason other than price. Three of those are the Game Changers, and removing them is what takes the floor away entirely rather than lowering it. Esper Sentinel is the most expensive nonland card in the core list at close to sixty dollars, and Midnight Clock takes the slot as a rock that refills your hand later instead of a tax nobody at a bracket 2 table was going to pay anyway. Uldaros Theorix comes out because a preorder legend is the wrong ten dollars to be spending on a list at this price, and Cloudreader Sphinx keeps the Sphinx count where it was.

Going up to bracket 4 starts with a land, again

Ancient Tomb at $127.05 is the single most expensive card added and the most relevant one on the page, because a deck whose payoff costs nine mana wants two colourless on turn one more than it wants another spell. Mana Vault, Rhystic Study, Mystic Remora and The One Ring are the ordinary additions at this level. Grand Arbiter Augustin IV is the one that is specific to this commander: it reduces white spells by and blue spells by as two separate effects, and a white, blue and black spell is both, so it takes off a cost that three other cards in the list are already chewing on.

The rung runs twelve Game Changers, which is a long way past bracket 3 and is the point of it. Worth noticing what is absent. Fierce Guardianship is the free counterspell every Esper deck at this power level plays, and it asks you to control a commander who is deliberately sitting in the command zone, so Force of Negation does that job here and charges a card for it. The cuts are the slow half of the core list: two taxing enchantments that a bracket 4 table walks through, two rocks that arrive too late, and four of the biggest Sphinxes, which takes the average mana value from 3.92 down to 3.49.

FAQ

Does The Ur-Sphinx's eminence work from the command zone?

Yes. The ability is written to work while the card is in the command zone or on the battlefield, so every other Sphinx spell you cast costs one mana less from your first turn onward, whether or not you ever cast him. The attack trigger is a separate ability with no such wording, so that one needs him on the battlefield.

Do you have to cast The Ur-Sphinx to get value out of him?

No, and this build assumes there are games where you never do. The discount is the reliable half and it costs nothing. Casting him is what turns on the mill and the free spells, and that is a nine-mana project rather than a plan.

What colors is The Ur-Sphinx?

White, blue and black, which makes him an Esper commander. He costs six generic mana plus one white, one blue and one black, for a mana value of nine, and he is a 10/10 Sphinx Avatar with flying.

When is The Ur-Sphinx legal in Commander?

He becomes playable on 2 October 2026, the day Reality Fracture Commander reaches shelves. Before then his status reads as illegal, which here describes a calendar and not a ban, and the other unreleased cards in this build are in exactly the same position.

How much does a The Ur-Sphinx deck cost?

The core list on this page is at least $664.80, the budget rung at least $181.36 and the bracket 4 rung at least $1,279.12. Every card is counted at its cheapest paper printing, and each figure says at least because three cards in the list have no price yet.

How many Sphinxes should a The Ur-Sphinx deck run?

This build runs twenty-five in the ninety-nine, plus the commander. The number is set by how many are worth playing rather than by a target, because only two Sphinxes in these colors cost less than four mana and the tribe runs out of good cheap options long before it runs out of expensive ones.

What bracket is this deck?

Three Game Changers put the core list on a bracket 3 floor. Read that number as the lowest slot the deck can truthfully be entered in rather than as a score, and during preview season the only direction it can travel is up, since the Game Changers list gets redrawn when the brackets themselves are revised and not when a set ships.