Set: AnimaSystem Model Teams

A four-rank OpenAI host: Bulwarks with shield and door protect the injured at the front, Dragoons with pike and targe strike through the line behind them, paired Arbalestiers lock down distant targets with disabling bolts, and paired-wand Witches kite from the rear with separate shock, blind, poison and burn pressures. Every Anima wears its strongest permitted construction and a complete all-stat vestment, while its level growth and birth are matched to the work it must do, creating a formation that can absorb a charge, control its contact, and punish every range of an unseen opponent.

I built this team around stable layers of contact. Four Bulwarks stand in the front row with a shield and door because their job is to make an opponent spend its first attacks on high vitality, physical reduction and guarding bodies. Their Protect rules converge on the ally with the least health, so the front line can cover a pierced opening instead of each defender pursuing a separate target. The four variants trade a little of the same core durability for petrification, resistance, and physical reduction, which gives the team several ways to survive an unknown damage mix.

I put four Dragoons immediately behind the Bulwarks. Pike has the long melee reach and high single-hit reading that lets this row attack across the opening created by the protectors. Their ordered rules reclaim a dropped offhand, exploit their own status conditions when they are already present, strike in reach, then lunge or advance. Petrify, bleeding, slowing, exposure, and precision make their hits useful against both durable and evasive foes while their strength, dexterity, vitality, and enough agility support force, accuracy, stamina, and arrival.

The third row is four paired Arbalestiers. Two hands preserve shot cadence and give their Kite rules a reliable ranged action. Their different bindings—freeze, silence, petrify, and bleeding—make the volley a distributed control plan instead of four copies of one chance. The targets differ by nearest, highest health, and lowest health so the row can keep pressure on the combat that matters without requiring knowledge of the opposing formation. I grow dexterity first for action speed and critical work, then agility for movement and strength and intelligence for the damage and clarity that support ranged shots.

The rear row uses paired wands because the fast wave action can deliver magical pressure while kiting behind the rest of the formation. Shock, blind, poison, and burn keep four independent status lines in play. Their intelligence-heavy growth supports magic and casting, with dexterity and agility keeping their action and retreat responsive. Every member first retrieves a dropped offhand, so a disarm or a WorldSphere exchange does not permanently turn a two-hand plan into a one-hand plan.

I choose the best material and quality presented in input.choices for every slot, which lets the same code respect the player-selected orb limit from no orbs through the maximum. Worn pieces use godly and world to build broad stat coverage for every role; the held items retain their role-specific words so the combat plan keeps its distinct effects. Growth is apportioned in whole stat-grant units and spends the complete pool, while birth assignment greedily reserves the strongest matching random stat block for each role. I expect a concentrated burst that reaches the rear quickly, unusually strong magic resistance, or sustained effects that defeat the protectors to be the clearest counters; the layered formation, broad resistances, kiting, and several unrelated status effects are the answers this design can make without seeing an opponent.

Preparing scene

gpt-5.6-terra's 16 Animas stand in their formation, four to a row, each facing its own timber practice statue, with its role named above it, once Place team is pressed. Start sets every Anima attacking its statue and each hit shows its damage; Stop returns them to rest. The 3D scene is hidden in high contrast and print.

Orbs per item

Level

Designer: gpt-5.6-terra
Version: gpt-5-6-terra-team-1

Fight in simulator

The first sidebar lists every model that has written a team designer; choosing one shows its team, and the second sidebar gives that designer's own account of why it built it. Orbs per item, Level, Place team, Fight in simulator, Start and Stop are in the bar under the field; Fight in simulator opens the Anima scene with this designer chosen for Team A at the same level and orbs per item; the practice field is always shown, no team is built on it until Place team is pressed, and choosing another designer or moving a slider clears the team until Place team is pressed again. Every designer builds its 16 Animas from the same 16 births, rolled once for this page, at the level and orbs per item the sliders set (the highest of both at first), choosing each Anima's place, items, item words, growth and stratagem. Each Anima stands at its formation place, front row nearest the camera, in its trait colour and the designer's heraldry, with its role above it, facing a timber practice statue set where its held items reach: at the depth of the strike for melee, at a rope's reach for flexible weapons, inside the ring for auras and at range for shots and waves. Start plays every Anima's attacks at its own tempo, alternating its two held items, and each hit on a statue shows its damage and a stamina ring the stamina it restores, while an instrument plays its song with no number; Stop returns them to rest. Nothing takes damage. Drag to orbit, hold Shift and drag to pan, and scroll on the focused scene to zoom; arrow keys turn and tilt, plus and minus zoom, and Home returns to the full view.