Chaos BibleLunia Handbook
Englishen

Updated 10 September 2026 · evidence through 9 September, 22:13 UTC

Character tier list

Compare damage, protection and party value in Chaos Lunia PvE, with character explanations and clear limits.

Current Open Beta 1.2.8 balance, including its published corrections. Level 75+ is the main comparison; lower-level progression and high-gear conditions are explained separately. PvP is outside this list.

What changed since the first review

The checked live combat rules are unchanged. This review adds 4,381 qualifying recent runs, corrects buff-stacking assumptions and ordinary skill ranks, and gives a practical plan for every character. Broad support-only S ranges have been narrowed; they were not proof of S-tier personal damage.

Read a range as conditional usefulness, not a precise rank. A / S means that the stronger assessment needs the conditions in the explanation. ? means that a defensible final tier is still open, not that the character is weak. Characters within a band are not ranked against one another.

A range is placed under its lower letter. Open the character explanation for the conditions behind the upper letter.

Overall

A broad character choice across solo and party play, gear levels and learning effort. Versatility matters alongside peak performance.

Strong contribution

Dacy

A

A strong damage candidate when her dolls can keep working. The A assessment remains provisional because the independent sample is limited.

Why this assessment? Dacy

Kali

A

A versatile damage-and-support choice. The A assessment depends on useful attacks and support that allies actually benefit from.

Why this assessment? Kali

Useful with clear trade-offs

Eir

B / A

A recovery specialist whose value grows when a party needs healing and a second chance. Independent damage farming is a different job.

Why this assessment? Eir

Sieg

B / A

A flexible melee choice with personal defenses. Recent small-party and raid comparisons are mixed rather than a clear universal lead, so the range stays provisional.

Why this assessment? Sieg

Dainn

B

A ranged area specialist with useful control. He needs targets to stay in his attacks and does not have the same recovery role as a healer.

Why this assessment? Dainn

Tia

B / A

A mobile specialist whose party value cannot be read from her damage share. Coordination determines how much of her potential is realized.

Why this assessment? Tia

Lime

B / A

Strong form-based clearing, with a more demanding rhythm than a simple damage table suggests. Form timing and mobility shape the result.

Why this assessment? Lime

Krieg

B

A melee support option with meaningful recovery and magical protection. His theoretical transformation ceiling is not consistently reproduced by observed damage.

Why this assessment? Krieg

Yuki

B / A

A strong control-oriented choice with encouraging boss evidence. Her versatility comes from combining pressure with freezing, not from a healer's toolkit.

Why this assessment? Yuki

Arien

B / A

A boss specialist with a large gap between favorable single-target conditions and general clearing. Positioning matters as much as the bow's range.

Why this assessment? Arien

Dark Eir

B / A

A versatile damage-and-recovery choice with stronger recent small-party evidence than raid evidence. The new B / A range reflects that uncertainty rather than a gameplay nerf.

Why this assessment? Dark Eir

More demanding or less consistent

Ryan

C / B

A mobile clearing pick whose small-party and raid results differ sharply. Choose him for the encounter pattern, not one theoretical DPS number.

Why this assessment? Ryan

Arta

C / B

A setup-oriented engineer with useful control options. Current evidence supports a cautious range rather than a top general recommendation.

Why this assessment? Arta

Final tier open

Asuka

?

Final tier open. Her mobility and protected attacking deserve consideration, but the available comparable runs cannot establish a reliable overall placement.

Why this assessment? Asuka

Ralph

?

Final tier open. The matched sample is too concentrated to separate the character from the few players representing him.

Why this assessment? Ralph

Gaon

?

Final tier open. Taunts and personal protection give him a role, but the small matched sample cannot establish its total value.

Why this assessment? Gaon

Iris

?

Final tier open. Her support costs and target-dependent attacks need evaluation beyond the very small comparable sample.

Why this assessment? Iris

PvE stages

Completing whole stages: fighting, moving between packs, surviving and helping the party finish. A boss damage chart cannot describe this on its own.

Strong contribution

Lime

A

A strong observed contribution across combat-heavy stages. Whole-run value still depends on getting into position and handling time outside key forms.

Why this assessment? Lime

Dacy

A

Dolls can keep contributing while Dacy repositions. Setup, retargeting and summon survival determine how much of that advantage survives a real stage.

Why this assessment? Dacy

Kali

A

Useful damage plus party assistance fits mixed stages. Supporting and transforming still consume time and resources.

Why this assessment? Kali

Useful with clear trade-offs

Eir

B / A

Healing, cleansing and resurrection can turn an unstable run into a completion. The upper band assumes those tools are needed rather than mostly overhealing.

Why this assessment? Eir

Sieg

B / A

Self-protection and close-combat pressure suit mixed stages. Travel, target access and dangerous recovery frames still affect clear speed.

Why this assessment? Sieg

Tia

B / A

Mobility and fog control help solo routing; cooldown resets matter in parties. Low personal raid damage is not a complete account of her contribution.

Why this assessment? Tia

Krieg

B

Useful when healing and protection fill a party gap. Lower ordinary clearing pressure and form downtime keep the broad assessment conservative.

Why this assessment? Krieg

Yuki

B / A

Control can make repeated engagements more manageable. The benefit shrinks when enemies resist freezing or move out of her coverage.

Why this assessment? Yuki

Arien

B / A

Very productive when important bosses dominate the run. A stage with many scattered packs can favor a broader clearing kit.

Why this assessment? Arien

Ryan

B / A

Good wave pressure and movement can suit farming and small parties. Boss-heavy raids pull the assessment down.

Why this assessment? Ryan

More demanding or less consistent

Arta

C / B

Devices are useful when the fight stays near them. Moving stages and repeated setup can reduce their whole-run contribution.

Why this assessment? Arta

Final tier open

Asuka

?

Directional attacks, control resistance and a magical shield can make encounters manageable. A final whole-stage tier needs more comparable evidence.

Why this assessment? Asuka

Ralph

?

Close-range combinations and movement offer a distinct playstyle, but there is not enough comparable whole-stage evidence for a final tier.

Why this assessment? Ralph

Gaon

?

Reach, control and threat tools can help in suitable stages. A final ranking needs more comparable solo and party runs.

Why this assessment? Gaon

Iris

?

She combines chain-sword attacks with party resource tools. There is insufficient comparable evidence to judge overall stage value fairly.

Why this assessment? Iris

Bossing

Useful contribution against bosses: sustained and burst damage, safe attacking opportunities, recovery and party support. The upper end of a support range assumes that the party needs and uses that support.

Strong contribution

Dacy

A / S

Persistent doll damage can be very strong against an accessible target. Movement, failed contact and lost summons reduce the ceiling.

Why this assessment? Dacy

Useful with clear trade-offs

Eir

B / A

Useful recovery and protection can outweigh modest personal damage when the group needs them. The B / A range includes revive restrictions and shield downtime; it is not a promise that every learning party gains an S-tier healer.

Why this assessment? Eir

Sieg

B / A

Good pressure when melee contact is reliable. Short protected attacks and defensive buffs help, but cannot replace moving out of immunity-piercing mechanics.

Why this assessment? Sieg

Dainn

B

Burst and repeated spell hits need a suitable opening. A spell's maximum hit count is not a promise against a small or moving boss.

Why this assessment? Dainn

Tia

B / A

Personal boss damage remains below the matched peers in the new sample. Timed cooldown resets can lift her party contribution, but limited targets, readiness and positioning prevent a blanket S assessment.

Why this assessment? Tia

Lime

B / A

Good when a form can maintain contact and land its repeated attacks. Movement, target shape and form downtime prevent a universal top boss rating.

Why this assessment? Lime

Krieg

B / A

His upper band needs a party benefiting from magic protection and enough safe melee time. It does not apply to physical or immunity-piercing boss attacks.

Why this assessment? Krieg

Yuki

B / A

Recent boss-target comparisons remain useful, but the earlier broad raid advantage is not stable across the new sample and alternative matching. Control and placement still decide how much of the kit works.

Why this assessment? Yuki

Arien

B / A

Overlapping Lightning Arrow hits can be exceptional against a large, reachable target. The upper band requires the correct distance and repeated hits; it is not a ranged safety guarantee.

Why this assessment? Arien

Kali

B / A

Good support potential during planned attack windows, with competitive rather than universally dominant own damage in the current comparisons.

Why this assessment? Kali

Dark Eir

B / A

Healing and planned protection add value beyond her own boss share. Fields need contact, and offensive blessings have a health cost over time.

Why this assessment? Dark Eir

More demanding or less consistent

Ryan

C

Recent raid and boss damage comparisons are less convincing than his small-party clearing. Repeated fields need target contact, and brief slide protection cannot cover every dangerous opening.

Why this assessment? Ryan

Arta

C / B

Sustained contact can reward prepared devices and transformation windows. A moving boss or lost setup makes the ceiling harder to realize.

Why this assessment? Arta

Final tier open

Asuka

?

Protection can preserve attacking opportunities against the right damage types. Blade Seal does not grant permanent immunity to all boss damage.

Why this assessment? Asuka

Ralph

?

Committed attacks need contact and a safe opening. His protected ultimate is a burst and safety tool, not permanent sustained output.

Why this assessment? Ralph

Gaon

?

A taunt matters only if the encounter responds to it. Protected attacks provide opportunities, not a guarantee against special boss mechanics.

Why this assessment? Gaon

Iris

?

Burst opportunities and a protected ultimate exist, but their practical repeatability and team value are not established by the current sample.

Why this assessment? Iris

Mobbing

Clearing groups: coverage, grouping and control, travel time, repeated attacks and reliability. Control and air combos lose value against enemies that resist them.

Strong contribution

Lime

A / S

Wide form attacks can be exceptional against packed enemies. The S end assumes good grouping and useful form uptime.

Why this assessment? Lime

Dacy

A

Dolls and control handle grouped combat well. Travel and target switching make peak boss-style output a poor direct estimate of farming speed.

Why this assessment? Dacy

Useful with clear trade-offs

Eir

B

Area attacks and control contribute, but frequent healing interrupts her own clearing. A party that already survives easily gains less from the extra recovery.

Why this assessment? Eir

Sieg

B

Launches and sword coverage handle nearby groups. Widely separated targets reduce the benefit, and his own survival tools are not party-wide healing.

Why this assessment? Sieg

Dainn

B / A

Area spells, roots, freezing and Ice Barrier can shape a pack. The upper band assumes controllable, grouped mobs and good placement.

Why this assessment? Dainn

Tia

B / A

Fog, blindness and fast movement help against suitable packs. Control-resistant mobs and uncoordinated positioning narrow the advantage.

Why this assessment? Tia

Yuki

B / A

Several freezing and slowing tools can make clustered mobs easier to handle. Fast scattered waves give less time to exploit that control.

Why this assessment? Yuki

Ryan

B / A

Repeated shots and fields can be strong when mobs stay in their coverage. Movement and grouped targets are important to the upper band.

Why this assessment? Ryan

Kali

B / A

Area pressure and control help clustered encounters. Chasing scattered mobs and long setup reduce the benefit.

Why this assessment? Kali

Arta

B

Decoys, grouping devices and hammer effects can organize eligible packs. They need placement and cannot be assumed to control every enemy.

Why this assessment? Arta

More demanding or less consistent

Arien

C / B

Coverage and retargeting are less convincing than the favorable boss case. Closely aligned packs help, while spread-out mobs reduce value.

Why this assessment? Arien

Final tier open

Asuka

?

Directional attacks and air combos can work well against eligible mobs. Targets that resist launches or are badly spaced change the result.

Why this assessment? Asuka

Ralph

?

Line attacks and air combinations can reward grouping. Long follow-throughs and spread-out mobs can reduce clearing efficiency.

Why this assessment? Ralph

Gaon

?

Spear coverage and grouping through threat can help eligible mobs. Spread and control resistance remain important limits.

Why this assessment? Gaon

Iris

?

Launches and area attacks can benefit from grouped, controllable enemies. These conditions are not guaranteed in every stage.

Why this assessment? Iris

Character explanations

EirStrongly scenario-dependent

A recovery specialist whose value grows when a party needs healing and a second chance. Independent damage farming is a different job.

Why this placement

B / A overall values her recovery role without pretending that healing is extra personal DPS. The new small-party damage sample is near the matched peers, while raids still show substantial ally healing. She rises when mistakes can be recovered; she gains less in already-safe speed farming. The previous B / S boss range overstated what the evidence could establish.

Practical plan

Solo
Use an offensive blessing when clearing safely, and heal after creating space. Her attack and defense blessings compete with each other; do not budget both bonuses at once.
Bosses
Plan Guard of Light around a predictable attack phase. It refreshes nearby protection while maintained; canceling the channel or leaving its area is not full-duration protection.
Mobs
Group enemies before committing an area attack, then reposition for Light of Healing. A heal cast has real recovery time; reacting only after repeated hits can trap you in a healing loop.
Party
Move close and face a fallen ally or tombstone before Resurrection. Stage revive limits still apply. Use Light Shield for a temporary recovery window; attacks and skill use end that shield.

Common mistake: Treating every protection effect as stackable, permanent attacking immunity. Ordinary shields, renewed guards and revive availability have different conditions.

Skills to understand

References at the ordinary learned rank cap, not a complete skill-point build. Bonus ranks can change costs and effects. Cooldowns are before applicable reductions. * marks skills with different cooldown-reduction rules. MP/HP below are casting costs; ongoing costs are explained in the tips. Quoted action times assume normal animation speed; speed changes and cancels alter the actual window.

SkillRankLevelRebirthBase cooldownCast MPCast HP
Light of Healing1172012 s4680
Resurrection552090 s1960
Light Shield4600150 s1900
Guard of Light5951120 s1,6800
Party contribution
Party heals, cleansing, resurrection and blessings. Light Shield ends on offensive action or skill use; Guard of Light provides a separate renewed protection window.
Gear and progression
Useful support exists before expensive damage gear. Heal scaling, learned ranks and MP management still matter; surplus healing does not speed up a safe run.
Handling and safety
Watch allies and incoming mechanics while positioning heals. Long channels and the decision to heal instead of attack are real costs.
What remains uncertain
No controlled measurement assigns an exact clear-time gain to each rescue or revive.
Recent combat evidence

Recent level-75+ sample: 553 appearances from 14 characters. Matched small parties: 10 characters / 200 comparisons. Matched raids: 7 characters / 122 comparisons.

These are observed personal-damage comparisons, not overall tier scores. 1.00 means parity with the available matched teammate after the Strength adjustment. The interval describes character-resampling uncertainty. It does not measure the damage enabled by a reset, prevented deaths or all regeneration effects. Cells below the minimum sample are withheld. The minimum is five characters, 30 paired comparisons and three encounter settings in that cell.

ContextDamage comparison95% interval
Small parties1.040.69 to 1.72
Raids0.820.66 to 1.07
Boss targets0.860.53 to 1.47
Mob targets1.241.05 to 2.66
Character and skills: Eir
SiegProvisional evidence

A flexible melee choice with personal defenses. Recent small-party and raid comparisons are mixed rather than a clear universal lead, so the range stays provisional.

Why this placement

B / A reflects usable melee pressure plus personal safety, not a proven raid damage lead. His recent solo clear comparisons are broadly typical. He fits encounters where he can stay close and finish attacks; scattered targets and forced disengagement make that harder. Party utility is narrower than Eir's or Tia's.

Practical plan

Solo
Choose Protective Aura for control resistance and regeneration or Magic Defense for the covered magical damage. These buffs compete, so verify which one is active instead of assuming both apply.
Bosses
Destruction Fist has a short protected section inside its attack, not immunity from start to finish. Time the contact and recovery around the boss's opening; Self-Sacrificing Blow provides a different protected commitment.
Mobs
Use launches and sword coverage to keep susceptible enemies together. Chasing a single displaced target can waste more time than finishing the remaining pack.
Party
Use threat control only when the enemies respond and you can handle their attention. Keep the boss or pack accessible to your allies rather than dragging it away from their attacks.

Common mistake: Assuming a ranged or glowing enemy attack is magical, or that interruption resistance prevents damage. Physical and neutral hits remain a different problem.

Skills to understand

References at the ordinary learned rank cap, not a complete skill-point build. Bonus ranks can change costs and effects. Cooldowns are before applicable reductions. * marks skills with different cooldown-reduction rules. MP/HP below are casting costs; ongoing costs are explained in the tips. Quoted action times assume normal animation speed; speed changes and cancels alter the actual window.

SkillRankLevelRebirthBase cooldownCast MPCast HP
Destruction Fist1173025 s2350
Protective Aura4590100 s690
Magic Defense5430120 s1660
Self-Sacrificing Blow5951120 s1,8100
Party contribution
Provocation and control can help where enemies respond to them. Most of his protection and damage buffs are personal.
Gear and progression
Strength and physical investment matter in the published balance. A strong defensive kit does not make a low-investment build equal to a developed one.
Handling and safety
Stay in reach, time protected hits and leave unsafe recovery windows. Melee commitment is the main trade-off.
What remains uncertain
Recent party and raid comparisons are mixed; a universal lead over other melee characters is not established.
Recent combat evidence

Recent level-75+ sample: 553 appearances from 22 characters. Matched small parties: 12 characters / 114 comparisons. Matched raids: 14 characters / 113 comparisons.

These are observed personal-damage comparisons, not overall tier scores. 1.00 means parity with the available matched teammate after the Strength adjustment. The interval describes character-resampling uncertainty. It does not measure the damage enabled by a reset, prevented deaths or all regeneration effects. Cells below the minimum sample are withheld. The minimum is five characters, 30 paired comparisons and three encounter settings in that cell.

ContextDamage comparison95% interval
Small parties0.970.57 to 2.21
Raids1.040.36 to 1.43
Boss targets1.460.99 to 2.23
Mob targets0.940.61 to 1.26
Character and skills: Sieg
DainnProvisional evidence

A ranged area specialist with useful control. He needs targets to stay in his attacks and does not have the same recovery role as a healer.

Why this placement

The revised B assessment replaces the earlier overly weak general impression. Recent matched solo clearing is near the cohort baseline, and mob damage is encouraging in both comparison methods. The small-party sample is still thin. Area pressure is a real strength, but long actions and lost target contact limit consistency.

Practical plan

Solo
Create room with Ice Barrier before a longer cast. Avoid placing a barrier where it spoils your own line of attack or makes you spend time walking around the fight.
Bosses
Use True Meteor Summons or Ultm. Crimson Judgement after a predictable commitment from the boss. Do not count every part of a large spell as a hit on a small or moving target.
Mobs
Place Rain of Fire where the pack will remain. Its interruption resistance is not damage immunity, and leaving the action early can change how much of the attack is delivered.
Party
Coordinate control and placement with the group's area attacks. Mana and resistance support help in suitable encounters, but do not replace direct recovery or revival.

Common mistake: Using a stationary full-hit calculation as actual DPS, or calling the class weak simply because one earlier group sample was unfavorable.

Skills to understand

References at the ordinary learned rank cap, not a complete skill-point build. Bonus ranks can change costs and effects. Cooldowns are before applicable reductions. * marks skills with different cooldown-reduction rules. MP/HP below are casting costs; ongoing costs are explained in the tips. Quoted action times assume normal animation speed; speed changes and cancels alter the actual window.

SkillRankLevelRebirthBase cooldownCast MPCast HP
Rain of Fire974025 s3920
Ice Barrier323025 s670
True Meteor Summons574150 s7430
Ultm. Crimson Judgement
Unlock required
1851150 s3,1000
Party contribution
Mana and elemental-resistance buffs, blocking terrain and crowd control. These help in the right encounter, without replacing healing or revival.
Gear and progression
Mana support is more noticeable while progressing. At high gear, actual hit coverage and casting opportunities remain important.
Handling and safety
Place fields ahead of movement and distinguish interruption resistance from damage immunity during long casts.
What remains uncertain
The raid comparison is carried by too few independent characters to prove a strong general boss rank.
Recent combat evidence

Recent level-75+ sample: 217 appearances from 8 characters. Matched small parties: 3 characters / 58 comparisons. Matched raids: 5 characters / 175 comparisons.

These are observed personal-damage comparisons, not overall tier scores. 1.00 means parity with the available matched teammate after the Strength adjustment. The interval describes character-resampling uncertainty. It does not measure the damage enabled by a reset, prevented deaths or all regeneration effects. Cells below the minimum sample are withheld. The minimum is five characters, 30 paired comparisons and three encounter settings in that cell.

ContextDamage comparison95% interval
Small partiesInsufficient comparisons
Raids1.400.59 to 3.65
Boss targets1.020.49 to 4.40
Mob targets1.321.20 to 1.75
Character and skills: Dainn
TiaStrongly scenario-dependent

A mobile specialist whose party value cannot be read from her damage share. Coordination determines how much of her potential is realized.

Why this placement

B / A includes mobility, control and cooldown support while acknowledging lower personal group damage. Recent solo clear comparisons are broadly typical. The old C / S boss span mixed a low damage reading with an unmeasured best-case support claim; the new assessment explains the useful party conditions directly.

Practical plan

Solo
Use Dark Fog for its attack and applicable blindness, with Spanning Throwing Knife as another tool rather than assuming fog is the whole build. Avoid spending the rotation only on a single sampled contributor.
Bosses
Turn Back Time does not damage the boss itself. Use it after allies have spent valuable skills and are able to cast again; it cannot erase their ongoing animation or positioning requirements.
Mobs
Use movement and fog to manage eligible packs. Keep enemies in allies' coverage; launches and control do not work equally on every target.
Party
At the ordinary rank cap, Turn Back Time reaches up to four nearby teammates. It excludes you and other copies of itself. Crowded raids need positioning, and String of Rescue is a situational rescue tool for an airborne ally.

Common mistake: Expecting the reset to pick the highest-damage allies automatically, reset all nine raid partners, or make already-ready skills more useful. Bonus ranks must not be confused with ordinary ranks.

Skills to understand

References at the ordinary learned rank cap, not a complete skill-point build. Bonus ranks can change costs and effects. Cooldowns are before applicable reductions. * marks skills with different cooldown-reduction rules. MP/HP below are casting costs; ongoing costs are explained in the tips. Quoted action times assume normal animation speed; speed changes and cancels alter the actual window.

SkillRankLevelRebirthBase cooldownCast MPCast HP
Dark Fog768050 s2770
Turn Back Time4310120 * s2660
String of Rescue972031 * s1090
Spanning Throwing Knife1073031 s880
Party contribution
Turn Back Time resets eligible cooldowns for nearby teammates, not Tia herself or another Turn Back Time. It adds little when useful skills are already ready.
Gear and progression
Cooldown reduction changes her own rotation. Teammates who already recast most skills quickly leave fewer valuable cooldowns to reset.
Handling and safety
Know which allies have spent their important skills, get them in range and time the reset. This is a high-coordination role.
What remains uncertain
Extra team damage from resets is credited to the teammates. Its causal size has not been measured in controlled runs.
Recent combat evidence

Recent level-75+ sample: 503 appearances from 32 characters. Matched small parties: 10 characters / 152 comparisons. Matched raids: 20 characters / 141 comparisons.

These are observed personal-damage comparisons, not overall tier scores. 1.00 means parity with the available matched teammate after the Strength adjustment. The interval describes character-resampling uncertainty. It does not measure the damage enabled by a reset, prevented deaths or all regeneration effects. Cells below the minimum sample are withheld. The minimum is five characters, 30 paired comparisons and three encounter settings in that cell.

ContextDamage comparison95% interval
Small parties0.800.29 to 0.91
Raids0.430.23 to 1.15
Boss targets0.650.27 to 0.94
Mob targets0.460.23 to 0.83
Character and skills: Tia
LimeProvisional evidence

Strong form-based clearing, with a more demanding rhythm than a simple damage table suggests. Form timing and mobility shape the result.

Why this placement

His strong mob range comes from form coverage and repeated contact, supported by the new group damage observations. Overall remains lower than that specialist ceiling because travel, form timing and target access matter. Recent solo speed does not prove that the largest damage totals always produce the fastest complete run.

Practical plan

Solo
Plan Bumba Lime and Cool Pang Pang Lime around the next useful fight rather than transforming immediately before a long walk. Include the time spent out of your chosen form in your expectations.
Bosses
Choose a form whose attacks can actually maintain contact with this boss. Large attacks do not guarantee that every repeated hit reaches a moving or oddly shaped target.
Mobs
Let susceptible mobs gather before committing the wide form attacks. Face the pack and keep a safe exit instead of wasting the form chasing a scattered enemy.
Party
Place recovery pickups from Rejuvenation Potion and Mana Potion where allies can reach them. They are not automatic remote healing, and recovery locations should not lure allies into a dangerous attack.

Common mistake: Treating a transformation's brief entry protection as immunity throughout the form, or evaluating only an uninterrupted stationary form loop.

Skills to understand

References at the ordinary learned rank cap, not a complete skill-point build. Bonus ranks can change costs and effects. Cooldowns are before applicable reductions. * marks skills with different cooldown-reduction rules. MP/HP below are casting costs; ongoing costs are explained in the tips. Quoted action times assume normal animation speed; speed changes and cancels alter the actual window.

SkillRankLevelRebirthBase cooldownCast MPCast HP
Bumba Lime1175060 s4920
Cool Pang Pang Lime574060 s4850
Rejuvenation Potion643045 s1330
Mana Potion643045 s1430
Party contribution
HP and MP pickups and selected buffs add assistance, while roots and freezing help eligible mobs. Pickups require access and are not automatic party-wide recovery.
Gear and progression
Physical investment and unlocked forms change his damage pattern. A theoretical uninterrupted form loop overstates what every encounter permits.
Handling and safety
Plan transformations around waves and travel. Large attacks still need correct facing, range and safe exits.
What remains uncertain
Small-party matching is thin, and observed solo failures include difficult content choices rather than just class weakness.
Recent combat evidence

Recent level-75+ sample: 267 appearances from 19 characters. Matched small parties: 5 characters / 44 comparisons. Matched raids: 11 characters / 149 comparisons.

These are observed personal-damage comparisons, not overall tier scores. 1.00 means parity with the available matched teammate after the Strength adjustment. The interval describes character-resampling uncertainty. It does not measure the damage enabled by a reset, prevented deaths or all regeneration effects. Cells below the minimum sample are withheld. The minimum is five characters, 30 paired comparisons and three encounter settings in that cell.

ContextDamage comparison95% interval
Small parties1.400.24 to 21.27
Raids1.250.63 to 3.21
Boss targets1.250.52 to 2.49
Mob targets1.740.90 to 3.15
Character and skills: Lime
DacyProvisional evidence

A strong damage candidate when her dolls can keep working. The A assessment remains provisional because the independent sample is limited.

Why this placement

A overall and A / S bossing recognize a leading sustained-damage candidate without claiming universal superiority. Her new party and raid damage remains strong, but solo clear speed is not a matching universal lead. Summon access and upkeep explain why the boss ceiling does not automatically become S in every category.

Practical plan

Solo
Establish Pandano when it can remain useful, then reposition without abandoning its target access. Summoning and retargeting are part of the time cost, not free background damage.
Bosses
Check that the dolls are making contact before committing Pandano's Dance. Movement phases, lost summons and unreachable targets can remove the apparent advantage even with strong equipment.
Mobs
Use Bundle Bundle Net against eligible packs to help maintain coverage. Do not scatter enemies beyond the dolls' useful attack area and then judge their damage by the stationary ceiling.
Party
Bring persistent damage while allies handle recovery and coordinated buffs. Happy Dance offers a protected action, but its timing and cooldown do not turn the whole rotation into permanent safety.

Common mistake: Assuming every generic summon damage entry identifies one particular doll, or that additional personal critical multiplier bypasses the dolls' own critical limit. The checked critical-damage multiplier cap for her dolls is 2.6×.

Skills to understand

References at the ordinary learned rank cap, not a complete skill-point build. Bonus ranks can change costs and effects. Cooldowns are before applicable reductions. * marks skills with different cooldown-reduction rules. MP/HP below are casting costs; ongoing costs are explained in the tips. Quoted action times assume normal animation speed; speed changes and cancels alter the actual window.

SkillRankLevelRebirthBase cooldownCast MPCast HP
Pandano2600120 s3050
Happy Dance6740150 s3820
Bundle Bundle Net962030 s1340
Pandano's Dance5951120 s2,0600
Party contribution
Binding, freezing and aggro management supplement the dolls. Her damage while moving is a personal reliability benefit, not a separate party buff.
Gear and progression
Dolls benefit from their owner's combat investment under the published rules, with a limit on their critical multiplier. Gear helps but does not remove summon behavior constraints.
Handling and safety
Manage summons, keep targets reachable and choose safe repositioning. Do not assume every doll attack lands continuously.
What remains uncertain
A small number of well-developed characters supplies much of the matched damage evidence. The exact lead remains uncertain.
Recent combat evidence

Recent level-75+ sample: 270 appearances from 9 characters. Matched small parties: 7 characters / 155 comparisons. Matched raids: 8 characters / 151 comparisons.

These are observed personal-damage comparisons, not overall tier scores. 1.00 means parity with the available matched teammate after the Strength adjustment. The interval describes character-resampling uncertainty. It does not measure the damage enabled by a reset, prevented deaths or all regeneration effects. Cells below the minimum sample are withheld. The minimum is five characters, 30 paired comparisons and three encounter settings in that cell.

ContextDamage comparison95% interval
Small parties1.700.76 to 2.03
Raids1.980.67 to 2.53
Boss targets1.510.77 to 3.02
Mob targets1.760.55 to 2.40
Character and skills: Dacy
KriegStrongly scenario-dependent

A melee support option with meaningful recovery and magical protection. His theoretical transformation ceiling is not consistently reproduced by observed damage.

Why this placement

The revised B general assessment gives proper weight to observed ally healing and the verified protection toolkit. Personal damage remains less convincing than the leading damage specialists, but the new raids record meaningful recovery. Bossing rises when that recovery and magic protection fit the encounter; it is not a hidden damage multiplier.

Practical plan

Solo
Balance form pressure from Vehement Transformation with safe opportunities to recover. Do not spend your remaining health on a support action while an incoming hit will finish you.
Bosses
Time Party Magic Defense for attacks of the covered magical types. It does not answer physical, neutral or immunity-piercing attacks, so first identify what is actually hurting the party.
Mobs
Keep nearby enemies within hammer or form coverage. Movement and setup between packs are why a favorable form ceiling is not a guaranteed fast farming result.
Party
Party HP Recovery and Party MP Recovery cost your own HP as well as MP. Use them when allies need the resource and you can afford the trade; coordinate recovery instead of repeatedly overhealing.

Common mistake: Calling him only a weak damage dealer, or treating party magic protection as universal invulnerability. Both ignore a different part of his actual role.

Skills to understand

References at the ordinary learned rank cap, not a complete skill-point build. Bonus ranks can change costs and effects. Cooldowns are before applicable reductions. * marks skills with different cooldown-reduction rules. MP/HP below are casting costs; ongoing costs are explained in the tips. Quoted action times assume normal animation speed; speed changes and cancels alter the actual window.

SkillRankLevelRebirthBase cooldownCast MPCast HP
Party HP Recovery1180030 s200400
Party MP Recovery968030 s270270
Party Magic Defense871030 s2340
Vehement Transformation675090 s3720
Party contribution
Party HP and MP recovery, vitality support and magical damage immunity. Recovery skills can cost his own HP as well as MP.
Gear and progression
Transformations and physical investment matter. Support costs and safe melee contact remain relevant at every gear level.
Handling and safety
Balance healing costs against personal safety, and align forms with useful attack windows instead of assuming perfect uptime.
What remains uncertain
Gear-matching choices move his raid result noticeably; a precise placement against nearby classes is not supported.
Recent combat evidence

Recent level-75+ sample: 203 appearances from 9 characters. Matched small parties: 7 characters / 62 comparisons. Matched raids: 7 characters / 81 comparisons.

These are observed personal-damage comparisons, not overall tier scores. 1.00 means parity with the available matched teammate after the Strength adjustment. The interval describes character-resampling uncertainty. It does not measure the damage enabled by a reset, prevented deaths or all regeneration effects. Cells below the minimum sample are withheld. The minimum is five characters, 30 paired comparisons and three encounter settings in that cell.

ContextDamage comparison95% interval
Small parties0.940.21 to 1.23
Raids0.730.52 to 4.82
Boss targets0.820.19 to 1.86
Mob targets0.940.56 to 1.04
Character and skills: Krieg
YukiProvisional evidence

A strong control-oriented choice with encouraging boss evidence. Her versatility comes from combining pressure with freezing, not from a healer's toolkit.

Why this placement

B / A is a control-and-pressure recommendation, not the earlier suggestion of a consistent general raid lead. Recent boss-target results are useful, but full raid and mob comparisons vary with the matching method and cohort. Her upper band needs susceptible targets and effective placement.

Practical plan

Solo
Use Cold Nucleus to help establish control before committing a longer attack. Keep an escape route for enemies that resist or outlast the control.
Bosses
Breath of Severe Cold protects about 3.4 seconds of its 4.33-second action if allowed to run. The ending is different from the protected part; reserve Ultm. Snowstorm of Queen for a useful opening rather than assuming it can replace all other attacks.
Mobs
Place freezing and slowing effects where the pack will stay, then concentrate your area damage. Control has no added value if the target is immune or would already die immediately.
Party
Help allies keep enemies inside their attack zones. Cold Incandescence resists interruption for much of its channel, but it is not a damage-immune channel.

Common mistake: Equating every ice effect with guaranteed boss freezing, or treating one favorable raid sample as proof of the same advantage everywhere.

Skills to understand

References at the ordinary learned rank cap, not a complete skill-point build. Bonus ranks can change costs and effects. Cooldowns are before applicable reductions. * marks skills with different cooldown-reduction rules. MP/HP below are casting costs; ongoing costs are explained in the tips. Quoted action times assume normal animation speed; speed changes and cancels alter the actual window.

SkillRankLevelRebirthBase cooldownCast MPCast HP
Breath of Severe Cold1174050 s2410
Cold Nucleus1073025 s2370
Cold Incandescence972045 s2330
Ultm. Snowstorm of Queen
Unlock required
1851150 s4,0900
Party contribution
Freezing, roots, mana and resistance support. Control must actually affect the target before it counts as a party benefit.
Gear and progression
More damage and cooldown investment improve repetition; target eligibility still matters even with excellent gear.
Handling and safety
Place attacks carefully and time protected actions. Interruption resistance on a channel does not make it safe from damage.
What remains uncertain
Boss-target evidence and whole-raid evidence differ; neither proves superiority in every party composition.
Recent combat evidence

Recent level-75+ sample: 633 appearances from 12 characters. Matched small parties: 9 characters / 286 comparisons. Matched raids: 10 characters / 203 comparisons.

These are observed personal-damage comparisons, not overall tier scores. 1.00 means parity with the available matched teammate after the Strength adjustment. The interval describes character-resampling uncertainty. It does not measure the damage enabled by a reset, prevented deaths or all regeneration effects. Cells below the minimum sample are withheld. The minimum is five characters, 30 paired comparisons and three encounter settings in that cell.

ContextDamage comparison95% interval
Small parties1.060.70 to 1.38
Raids0.820.40 to 1.81
Boss targets1.230.89 to 1.39
Mob targets0.810.36 to 1.20
Character and skills: Yuki
ArienStrongly scenario-dependent

A boss specialist with a large gap between favorable single-target conditions and general clearing. Positioning matters as much as the bow's range.

Why this placement

The boss range is narrowed to B / A: favorable contact remains important, but the new sample is too mixed and the raid comparison too small for a confident S claim. Solo clearing is encouraging in a small cohort. Her weak point is carrying that single-target setup across scattered packs and moving bosses.

Practical plan

Solo
Use Lightning Arrow from a position that gets useful overlap without forcing an unsafe approach. Keep a separate plan for quickly reaching and finishing scattered mobs.
Bosses
Wait for a readable stationary opening before the high-contact attack. More projectiles or a large effect does not mean every part hits one target. Compare the boss's movement and shape, not only your displayed attack.
Mobs
Line up enemies before committing your stronger attacks. Moving the whole fight to a suitable line can help more than repeatedly firing at isolated targets.
Party
Use Single Blow of Light Spirit when the group can exploit the same attack window. Spirit's Anger has a health casting cost; Spirit's Guard provides regeneration support, not universal damage immunity.

Common mistake: Treating an attack-component cap as a cap on all damage, or assuming unlimited extra attack improves every skill equally. Strength, critical effects, cooldowns and contact still differ.

Skills to understand

References at the ordinary learned rank cap, not a complete skill-point build. Bonus ranks can change costs and effects. Cooldowns are before applicable reductions. * marks skills with different cooldown-reduction rules. MP/HP below are casting costs; ongoing costs are explained in the tips. Quoted action times assume normal animation speed; speed changes and cancels alter the actual window.

SkillRankLevelRebirthBase cooldownCast MPCast HP
Lightning Arrow1074022 s1690
Spirit's Anger7580120 * s0364
Spirit's Guard459060 * s690
Single Blow of Light Spirit5951120 * s1,8900
Party contribution
Some mana and regeneration support, but most of her value comes from landing her own attacks. Personal damage buffs are not counted as a second team bonus.
Gear and progression
Cooldown reduction and attack investment change her rotation considerably. Actual hit overlap still determines how much of that investment lands.
Handling and safety
Learn the distance at which important hits overlap and when approaching a boss is unsafe.
What remains uncertain
Matched boss samples are small. A large favorable signal supports specialist potential, not a proven universal first place.
Recent combat evidence

Recent level-75+ sample: 247 appearances from 10 characters. Matched small parties: 6 characters / 51 comparisons. Matched raids: 4 characters / 16 comparisons.

These are observed personal-damage comparisons, not overall tier scores. 1.00 means parity with the available matched teammate after the Strength adjustment. The interval describes character-resampling uncertainty. It does not measure the damage enabled by a reset, prevented deaths or all regeneration effects. Cells below the minimum sample are withheld. The minimum is five characters, 30 paired comparisons and three encounter settings in that cell.

ContextDamage comparison95% interval
Small parties0.380.15 to 2.80
RaidsInsufficient comparisons
Boss targets1.090.38 to 2.65
Mob targets0.530.18 to 1.01
Character and skills: Arien
RyanStrongly scenario-dependent

A mobile clearing pick whose small-party and raid results differ sharply. Choose him for the encounter pattern, not one theoretical DPS number.

Why this placement

The small-party/raid split is still the defining evidence. New small parties and solo cohorts support his clearing role, while raid and boss comparisons remain less convincing. That supports better mob/PvE potential than the broad overall reading, without pretending that a faster cast rate solves every boss problem.

Practical plan

Solo
Combine True Impartial Shot with purposeful repositioning. High cooldown reduction should create useful attacks, not more resource spending while the target is outside the hit area.
Bosses
Place Sulfur Rain for a phase when the boss will stay in contact. Right Side Slide Shot has only a short initial protection window; its whole slide is not damage-immune.
Mobs
Use repeated shots and areas on grouped targets before moving to the next pack. The benefit depends on coverage rather than the number of damage events alone.
Party
Leave allies a stable attack area instead of pulling enemies away from their fields. Ultm. Quanta Gun is a long protected commitment, so align it with a useful phase rather than spending it on a nearly finished target.

Common mistake: Explaining all raid weakness with boss armor, or calling every repeated field tick a guaranteed hit. Neither conclusion follows from the observed damage alone.

Skills to understand

References at the ordinary learned rank cap, not a complete skill-point build. Bonus ranks can change costs and effects. Cooldowns are before applicable reductions. * marks skills with different cooldown-reduction rules. MP/HP below are casting costs; ongoing costs are explained in the tips. Quoted action times assume normal animation speed; speed changes and cancels alter the actual window.

SkillRankLevelRebirthBase cooldownCast MPCast HP
True Impartial Shot574120 s1690
Sulfur Rain1073025 s2000
Right Side Slide Shot1171013 s1610
Ultm. Quanta Gun
Unlock required
1851150 s3,4000
Party contribution
Movement, short protected slides and selected freezing effects. He does not bring the same direct recovery or cooldown-reset role as dedicated supports.
Gear and progression
High cooldown reduction allows more frequent casts, but does not make each field hit a moving target or erase resource costs.
Handling and safety
Reposition without losing field contact and avoid confusing many damage events with high effective boss damage.
What remains uncertain
The reason for every raid gap has not been isolated in a controlled test. Boss armor is not assumed as a blanket explanation.
Recent combat evidence

Recent level-75+ sample: 381 appearances from 17 characters. Matched small parties: 7 characters / 76 comparisons. Matched raids: 5 characters / 33 comparisons.

These are observed personal-damage comparisons, not overall tier scores. 1.00 means parity with the available matched teammate after the Strength adjustment. The interval describes character-resampling uncertainty. It does not measure the damage enabled by a reset, prevented deaths or all regeneration effects. Cells below the minimum sample are withheld. The minimum is five characters, 30 paired comparisons and three encounter settings in that cell.

ContextDamage comparison95% interval
Small parties1.600.61 to 4.21
Raids0.590.34 to 6.03
Boss targets0.550.27 to 1.44
Mob targets1.600.68 to 3.40
Character and skills: Ryan
KaliStrongly scenario-dependent

A versatile damage-and-support choice. The A assessment depends on useful attacks and support that allies actually benefit from.

Why this placement

A in general PvE reflects the new positive small-party and raid damage signals plus a verified support kit. It is still a provisional recommendation: target-specific results are mixed and solo clear speed is broadly typical. Her versatility comes from choosing the right support action, not adding every buff as free damage.

Practical plan

Solo
Keep an affordable damage loop and use the needed recovery between engagements. A transformation or support channel still costs time that belongs in the whole-run comparison.
Bosses
Use True Rhapsody of Blood and Epic of Hero when allies are ready to attack within range. The renewed guard is useful while maintained; overlapping support casts do not simply add their full durations together.
Mobs
True Song of Hero can help a grouped team while you pressure suitable packs. Avoid scattering targets or spending a long support setup after the pack is already effectively finished.
Party
Use Phantom Fantasia for a recovery opportunity when recipients can benefit. Its protective/regenerative bundle can end when the recipient attacks or uses a skill; coordinate the pause instead of promising uninterrupted protected DPS.

Common mistake: Quoting the strongest bonus-rank buff values as normal learned values, or judging her contribution only by the damage shown next to her own name.

Skills to understand

References at the ordinary learned rank cap, not a complete skill-point build. Bonus ranks can change costs and effects. Cooldowns are before applicable reductions. * marks skills with different cooldown-reduction rules. MP/HP below are casting costs; ongoing costs are explained in the tips. Quoted action times assume normal animation speed; speed changes and cancels alter the actual window.

SkillRankLevelRebirthBase cooldownCast MPCast HP
True Song of Hero1171150 s2310
True Rhapsody of Blood1171125 s2310
Phantom Fantasia1074090 s2420
Epic of Hero5951120 s1,6300
Party contribution
Party stat buffs, MP recovery, HP/MP regeneration and renewed protection through Epic of Hero. Some recovery protection ends when its recipient attacks or uses a skill.
Gear and progression
Flat buffs can be especially meaningful earlier; percentage bonuses and coordinated protection remain relevant later. Learned ranks change their strength.
Handling and safety
Choose when to attack, recover allies or channel protection. Team spacing and timing determine whether the support lands usefully.
What remains uncertain
No single damage-share figure captures these different support roles, and their exact team gain remains unmeasured.
Recent combat evidence

Recent level-75+ sample: 709 appearances from 25 characters. Matched small parties: 17 characters / 208 comparisons. Matched raids: 13 characters / 172 comparisons.

These are observed personal-damage comparisons, not overall tier scores. 1.00 means parity with the available matched teammate after the Strength adjustment. The interval describes character-resampling uncertainty. It does not measure the damage enabled by a reset, prevented deaths or all regeneration effects. Cells below the minimum sample are withheld. The minimum is five characters, 30 paired comparisons and three encounter settings in that cell.

ContextDamage comparison95% interval
Small parties1.320.62 to 1.67
Raids1.460.89 to 1.81
Boss targets0.860.63 to 1.11
Mob targets0.930.76 to 1.50
Character and skills: Kali
AsukaInsufficient comparable evidence

Final tier open. Her mobility and protected attacking deserve consideration, but the available comparable runs cannot establish a reliable overall placement.

Why this placement

The final tier stays open because the new like-for-like small-party and raid samples still contain only four characters each. Solo clearing looks broadly typical in a small cohort. Her real mobility and protection deserve analysis, but neither a low damage share nor the phrase 'cannot die with mana' establishes her overall tier.

Practical plan

Solo
Use Blade Seal's directional follow-up to keep contact and position well. Its sustained benefit is interruption resistance and movement; the opening damage immunity is only a short section.
Bosses
Veil of Swords has immunity through about 4.22 seconds, while its nominal action lasts 6 seconds and the listed dash cancel begins at 4.5. Plan for that gap and recovery instead of treating the full cast as safe.
Mobs
Exploit air combinations on launchable mobs, but keep targets where your follow-up can land. Unlaunchable enemies and scattered packs invalidate the same combo expectations.
Party
Choose Magic Defense or Protective Aura for the actual danger: the two buffs compete. Magical damage protection excludes physical/neutral damage, and these buffs use different cooldown rules from her fast-repeating damage skills.

Common mistake: Assuming zero cooldown on an attack gives zero cooldown on the shield, stacking both defensive buffs on paper, or confusing anti-interruption with all-damage immunity.

Skills to understand

References at the ordinary learned rank cap, not a complete skill-point build. Bonus ranks can change costs and effects. Cooldowns are before applicable reductions. * marks skills with different cooldown-reduction rules. MP/HP below are casting costs; ongoing costs are explained in the tips. Quoted action times assume normal animation speed; speed changes and cancels alter the actual window.

SkillRankLevelRebirthBase cooldownCast MPCast HP
Blade Seal1171025 s1900
Veil of Swords572090 s3840
Magic Defense5430120 * s1660
Protective Aura459060 * s690
Party contribution
Mostly personal protection and movement, plus control on eligible enemies. This can improve her usable output without becoming a party-wide damage buff.
Gear and progression
Cooldown reduction can sustain frequent eligible attacks. Magic Defense and Protective Aura use different cooldown rules; do not infer constant shield coverage from a fast attack rotation.
Handling and safety
Read the protection breakdown below. Direction choice, range and the distinction between damage immunity and interruption resistance matter.
What remains uncertain
The new period still has too few comparable characters in the relevant modes for a defensible final tier. The toolkit and advice remain useful, but a large result from one or two players is not a class verdict.
Recent combat evidence

Recent level-75+ sample: 99 appearances from 11 characters. Matched small parties: 4 characters / 16 comparisons. Matched raids: 4 characters / 11 comparisons.

These are observed personal-damage comparisons, not overall tier scores. 1.00 means parity with the available matched teammate after the Strength adjustment. The interval describes character-resampling uncertainty. It does not measure the damage enabled by a reset, prevented deaths or all regeneration effects. Cells below the minimum sample are withheld. The minimum is five characters, 30 paired comparisons and three encounter settings in that cell.

ContextDamage comparison95% interval
Small partiesInsufficient comparisons
RaidsInsufficient comparisons
Boss targetsInsufficient comparisons
Mob targetsInsufficient comparisons
Character and skills: Asuka

Asuka: protection is not one single effect

RalphInsufficient comparable evidence

Final tier open. The matched sample is too concentrated to separate the character from the few players representing him.

Why this placement

The final tier remains open. The recent matched raid result looks large, but it comes from only two characters; small-party matching has one. That is a player sample, not proof that Ralph has suddenly become a leading class. His actual attack commitments and positioning can still be described precisely.

Practical plan

Solo
Plan a safe exit before True Hurricane Surge Punch. Its nominal action is 3.5 seconds and its listed dash cancel starts at 2.8; it is not a quick, freely canceled poke.
Bosses
Aim the sequence through the boss's useful contact area. Multiple field placements are not automatically multiple hits on every boss. Ultm. Storm Demon Attack is a long protected action, not a measure of permanent sustained damage.
Mobs
Line up packs for the forward attack and use air combinations only on eligible mobs. Scattering enemies out of the follow-up makes impressive animations produce less useful clearing.
Party
Avoid moving the target away from allies' attacks. Protective Aura offers control resistance and regeneration; it is not a party-wide immunity skill.

Common mistake: Extrapolating a strong run by one geared player, or assuming the interruption protection on True Hurricane Surge Punch covers the whole action and also stops incoming damage.

Skills to understand

References at the ordinary learned rank cap, not a complete skill-point build. Bonus ranks can change costs and effects. Cooldowns are before applicable reductions. * marks skills with different cooldown-reduction rules. MP/HP below are casting costs; ongoing costs are explained in the tips. Quoted action times assume normal animation speed; speed changes and cancels alter the actual window.

SkillRankLevelRebirthBase cooldownCast MPCast HP
True Hurricane Surge Punch574120 s1600
Hurricane Massacre Kick572090 s1190
Protective Aura459060 s690
Ultm. Storm Demon Attack
Unlock required
1851150 s4,0100
Party contribution
Mostly personal movement, protective aura and aggro reduction. Air control helps only against eligible targets.
Gear and progression
Judge the kit at realistic physical investment. A perfect stationary rotation is not evidence of the damage delivered throughout a real encounter.
Handling and safety
Facing, close-range positioning and cancel timing demand attention; several important attacks have a substantial commitment.
What remains uncertain
The new period still has too few comparable characters in the relevant modes for a defensible final tier. The toolkit and advice remain useful, but a large result from one or two players is not a class verdict.
Recent combat evidence

Recent level-75+ sample: 36 appearances from 4 characters. Matched small parties: 1 characters / 2 comparisons. Matched raids: 2 characters / 9 comparisons.

These are observed personal-damage comparisons, not overall tier scores. 1.00 means parity with the available matched teammate after the Strength adjustment. The interval describes character-resampling uncertainty. It does not measure the damage enabled by a reset, prevented deaths or all regeneration effects. Cells below the minimum sample are withheld. The minimum is five characters, 30 paired comparisons and three encounter settings in that cell.

ContextDamage comparison95% interval
Small partiesInsufficient comparisons
RaidsInsufficient comparisons
Boss targetsInsufficient comparisons
Mob targetsInsufficient comparisons
Character and skills: Ralph
Dark EirProvisional evidence

A versatile damage-and-recovery choice with stronger recent small-party evidence than raid evidence. The new B / A range reflects that uncertainty rather than a gameplay nerf.

Why this placement

B / A replaces the fixed A recommendation because new small parties remain favorable while raid damage is less stable. Healing, blessings and guards still provide broad usefulness. This is a correction to assessment confidence with unchanged checked combat rules, not evidence that the character was nerfed.

Practical plan

Solo
Keep enough health to fund the kit's costs and use Shadow of Healing after creating a safe opening. Recovery that prevents a death can be worth more than forcing another attack.
Bosses
Use Guard of Darkness for a planned party window. Shield of Darkness is a different shield that ends with offensive action or skill use, so it cannot simply be counted as full-time attacking protection.
Mobs
Keep enemies in the areas that are already active rather than repeatedly setting up on moving targets. Reposition heals around where allies will finish dodging.
Party
Power of Darkness provides offensive support with an ongoing health cost. Check whether the group can sustain that cost and needs recovery. She cannot substitute Eir's resurrection when a revive is required.

Common mistake: Treating her curse-resistance reduction as a universal percentage damage-taken debuff, or treating every shield and guard as the same stackable effect.

Skills to understand

References at the ordinary learned rank cap, not a complete skill-point build. Bonus ranks can change costs and effects. Cooldowns are before applicable reductions. * marks skills with different cooldown-reduction rules. MP/HP below are casting costs; ongoing costs are explained in the tips. Quoted action times assume normal animation speed; speed changes and cancels alter the actual window.

SkillRankLevelRebirthBase cooldownCast MPCast HP
Shadow of Healing1075012 s2460
Power of Darkness10730120 s2390
Shield of Darkness9720110 s3510
Guard of Darkness5951120 s1,9100
Party contribution
Party healing, strength and attack blessings, sleep and protective tools. Shield of Darkness ends on offensive action or skill use; Guard of Darkness is a separate renewed effect. She has no Eir-style resurrection.
Gear and progression
Useful recovery does not require maximum damage investment. Buff ranks and the party's actual need for healing determine the return.
Handling and safety
Manage health costs and long actions while keeping allies within support range. Choose recovery before a party failure, not after it.
What remains uncertain
The all-round recommendation is a role judgment; no controlled test proves a fixed clear-time advantage over every specialist.
Recent combat evidence

Recent level-75+ sample: 393 appearances from 31 characters. Matched small parties: 9 characters / 67 comparisons. Matched raids: 17 characters / 97 comparisons.

These are observed personal-damage comparisons, not overall tier scores. 1.00 means parity with the available matched teammate after the Strength adjustment. The interval describes character-resampling uncertainty. It does not measure the damage enabled by a reset, prevented deaths or all regeneration effects. Cells below the minimum sample are withheld. The minimum is five characters, 30 paired comparisons and three encounter settings in that cell.

ContextDamage comparison95% interval
Small parties1.231.01 to 5.30
Raids0.570.33 to 1.10
Boss targets0.580.37 to 1.37
Mob targets1.040.51 to 1.54
Character and skills: Dark Eir
ArtaProvisional evidence

A setup-oriented engineer with useful control options. Current evidence supports a cautious range rather than a top general recommendation.

Why this placement

The cautious C / B range reflects below-peer direct comparisons that change substantially after accounting for which teammate classes were compared. That disagreement and the small raid sample prevent a confident high rank. Devices and control have real uses, but setup must pay off before the fight moves elsewhere.

Practical plan

Solo
Use Target Decoy F-4 and Magnetic Field Generator S-2 where the next fight will actually take place. Preplacing devices in an abandoned area turns setup time into lost progress.
Bosses
Commit Artitan Mk. 28 when the target is reachable for a useful portion of the form. A moving or untargetable phase can waste the setup even if the stationary ceiling looks good.
Mobs
Use decoys and grouping against responsive mobs, then keep attacks inside that area. The hammer's named freezing effect is a chance-based immobilization effect, not guaranteed full crowd control.
Party
Choose one appropriate hammer coating; they compete rather than stacking all their bonuses. Freezing Hammer does not guarantee a proc on every hit or control every boss.

Common mistake: Counting every device as permanently hitting, adding all hammer coatings together, or trusting a single comparison method when the alternative gives a different picture.

Skills to understand

References at the ordinary learned rank cap, not a complete skill-point build. Bonus ranks can change costs and effects. Cooldowns are before applicable reductions. * marks skills with different cooldown-reduction rules. MP/HP below are casting costs; ongoing costs are explained in the tips. Quoted action times assume normal animation speed; speed changes and cancels alter the actual window.

SkillRankLevelRebirthBase cooldownCast MPCast HP
Target Decoy F-41073023 s2390
Magnetic Field Generator S-21074025 s2910
Artitan Mk. 288740180 s7270
Freezing Hammer10740180 s2420
Party contribution
Local control and distraction rather than broad healing or party stat buffs. Device ownership is included in observed damage where recorded.
Gear and progression
Cooldown investment helps repeat actions, but cannot remove setup time or make a device reach every target.
Handling and safety
Plan the fight around device placement and transformation timing. This is a more deliberate style than immediate mobile clearing.
What remains uncertain
Only a few independent characters remain after matching; device behavior and thin samples make adjacent placements uncertain.
Recent combat evidence

Recent level-75+ sample: 246 appearances from 8 characters. Matched small parties: 5 characters / 119 comparisons. Matched raids: 4 characters / 95 comparisons.

These are observed personal-damage comparisons, not overall tier scores. 1.00 means parity with the available matched teammate after the Strength adjustment. The interval describes character-resampling uncertainty. It does not measure the damage enabled by a reset, prevented deaths or all regeneration effects. Cells below the minimum sample are withheld. The minimum is five characters, 30 paired comparisons and three encounter settings in that cell.

ContextDamage comparison95% interval
Small parties0.780.51 to 1.18
RaidsInsufficient comparisons
Boss targets0.570.34 to 0.90
Mob targets0.790.44 to 0.97
Character and skills: Arta
GaonInsufficient comparable evidence

Final tier open. Taunts and personal protection give him a role, but the small matched sample cannot establish its total value.

Why this placement

The final tier remains open: the recent small-party match has one character and the raid match four. Lower observed damage does not quantify the value of holding threat or making enemies easier to hit. The right conclusion is a conditional melee/control role with insufficient comparative evidence, not a fabricated precise rank.

Practical plan

Solo
Use reach and keep a safe recovery plan. Lightning Leap has a brief damage-immunity section inside a longer action; the whole leap is not invulnerable.
Bosses
Check that this boss responds to Provocation before relying on threat control. Self-Sacrificing Blow is a distinct protected commitment; immunity-piercing mechanics still need a movement plan.
Mobs
Bring responsive enemies into spear coverage without spreading them beyond the group's attacks. Wide reach is valuable only when the targets stay in it.
Party
Use Magic Defense for the covered magical damage, not as an answer to every attack. Taking attention helps only if you can survive it and improve your allies' attack opportunities.

Common mistake: Calling every taunt reliable tanking on every boss, or assigning a confident tier from one small-party representative.

Skills to understand

References at the ordinary learned rank cap, not a complete skill-point build. Bonus ranks can change costs and effects. Cooldowns are before applicable reductions. * marks skills with different cooldown-reduction rules. MP/HP below are casting costs; ongoing costs are explained in the tips. Quoted action times assume normal animation speed; speed changes and cancels alter the actual window.

SkillRankLevelRebirthBase cooldownCast MPCast HP
Provocation542030 s1430
Magic Defense6550180 s1730
Lightning Leap1075015 s3280
Self-Sacrificing Blow5951120 s1,8000
Party contribution
Provocation, control and personal magical protection. Drawing attention is useful only when it actually helps the party survive and attack.
Gear and progression
Physical investment, learned ranks and target access matter. A protected ultimate does not remove the demands of ordinary melee combat.
Handling and safety
Manage reach, facing and the danger of taking threat. Do not replace movement with a protection assumption.
What remains uncertain
The new period still has too few comparable characters in the relevant modes for a defensible final tier. The toolkit and advice remain useful, but a large result from one or two players is not a class verdict.
Recent combat evidence

Recent level-75+ sample: 42 appearances from 5 characters. Matched small parties: 1 characters / 9 comparisons. Matched raids: 4 characters / 43 comparisons.

These are observed personal-damage comparisons, not overall tier scores. 1.00 means parity with the available matched teammate after the Strength adjustment. The interval describes character-resampling uncertainty. It does not measure the damage enabled by a reset, prevented deaths or all regeneration effects. Cells below the minimum sample are withheld. The minimum is five characters, 30 paired comparisons and three encounter settings in that cell.

ContextDamage comparison95% interval
Small partiesInsufficient comparisons
RaidsInsufficient comparisons
Boss targetsInsufficient comparisons
Mob targetsInsufficient comparisons
Character and skills: Gaon
IrisInsufficient comparable evidence

Final tier open. Her support costs and target-dependent attacks need evaluation beyond the very small comparable sample.

Why this placement

The final tier stays open. The new sample now has some small-party comparisons, but only three characters there and two in raids. The important correction is to assess her actual area attack and support trade-offs, rather than keep an old one-skill damage assumption or automatically label her last.

Practical plan

Solo
Place Dance of Flames where targets will remain and keep resources for the next pack. Its brief interruption resistance is not protection for the entire area-attack action.
Bosses
Do not assume a launch or a full repeated area works on this boss. Preserve a safe exit and judge landed attacks, not the nominal number of effects on screen.
Mobs
Use launches and areas together on susceptible groups. Your useful damage changes when the pack cannot be lifted or leaves the placed area.
Party
Unholy Sacrifice, Spiritual Bargain and Deathly Grace compete instead of stacking all three effects. Choose the resource exchange or maximum-HP support the team needs; the exchange has ongoing costs despite a zero HP casting cost in the table. For Deathly Grace, ordinary rank 3 gives 15% maximum HP; 25% belongs to bonus rank 5.

Common mistake: Adding the three party buffs together, promising free healing, or quoting the bonus-rank maximum-HP value as the normal learned value.

Skills to understand

References at the ordinary learned rank cap, not a complete skill-point build. Bonus ranks can change costs and effects. Cooldowns are before applicable reductions. * marks skills with different cooldown-reduction rules. MP/HP below are casting costs; ongoing costs are explained in the tips. Quoted action times assume normal animation speed; speed changes and cancels alter the actual window.

SkillRankLevelRebirthBase cooldownCast MPCast HP
Dance of Flames1074040 s2080
Unholy Sacrifice7640120 * s3080
Spiritual Bargain7600120 * s1630
Deathly Grace3400120 * s1500
Party contribution
Party maximum-HP support and HP/MP exchange buffs. Exchanging one resource for another is a real cost, not free healing.
Gear and progression
Cooldown reduction, ranks and enough resources influence how comfortable the kit feels. A high theoretical cast rate alone does not establish its effectiveness.
Handling and safety
Watch both health and mana, place attacks and account for team resource costs before applying support.
What remains uncertain
The new period still has too few comparable characters in the relevant modes for a defensible final tier. The toolkit and advice remain useful, but a large result from one or two players is not a class verdict.
Recent combat evidence

Recent level-75+ sample: 147 appearances from 5 characters. Matched small parties: 3 characters / 5 comparisons. Matched raids: 2 characters / 9 comparisons.

These are observed personal-damage comparisons, not overall tier scores. 1.00 means parity with the available matched teammate after the Strength adjustment. The interval describes character-resampling uncertainty. It does not measure the damage enabled by a reset, prevented deaths or all regeneration effects. Cells below the minimum sample are withheld. The minimum is five characters, 30 paired comparisons and three encounter settings in that cell.

ContextDamage comparison95% interval
Small partiesInsufficient comparisons
RaidsInsufficient comparisons
Boss targetsInsufficient comparisons
Mob targetsInsufficient comparisons
Character and skills: Iris

How to read the list

Usefulness comes before damage rank

A revive, a well-timed cooldown reset and a personal damage buff solve different problems. Ally protection and recovery count as party contributions. Personal protection explains when damage can be delivered safely; it is not added again as invented extra DPS. A shield that ends when you attack is different from protection that lets you keep attacking.

Fair comparisons have limits

The combined review contains 9,691 qualifying runs from 395 characters. The new period contributes 4,381 runs from 292 characters; the earlier comparison contains 5,305 runs. Five runs crossing the cutoff are used only in the combined context. Main performance comparisons require level 75+, the same run and similar Strength scaling, with tighter equipment checks and a cross-check accounting for which teammate classes were compared. Repeated farming by one character is not counted as independent votes. Team composition, builds and skill still matter.

Weights express a playstyle

The balanced reading gives solo and party play equal attention, and bosses and mobs equal attention. It was challenged with 30/70 and 70/30 scenario mixes, tighter gear matching, separate small-party and raid views, and removal of dominant contributors. These are review priorities, not a formula that turns healing into damage. PvE judges whole-stage usefulness; Overall also asks whether that usefulness carries across gear and playstyles. The four categories are not added together.

Why some ranges are wide

More party reliance raises the value of Eir, Tia and protective support; more independent farming raises the value of mobile clearing. Large stationary bosses reward overlapping hits, while moving targets reduce repeated field hits. Better cooldown reduction changes rotations, but cannot remove animation time, positioning or resource costs. The range shows these changes instead of hiding them behind a single score.

What is verified, and what is judgment

Skill behavior and the published balance were checked against the game rules. Run comparisons are observations, not controlled team-swap experiments. The tier ranges are an editorial interpretation of both. No full combat simulation is treated as an in-game measurement, and there is not enough evidence to assign every character a final tier or to prove that neighboring bands differ.

Asuka: protection is not one single effect

Blade Seal is a major source of her observed damage. Its opening has damage immunity from 0.11 to 0.31 action seconds; its directional follow-up has interruption protection, not continuous damage immunity. At normal action speed the follow-up lasts about 2.27 seconds. Veil of Swords has a different immunity window, about 4.22 seconds of a 6-second action. Magic Defense protects against magical damage types, with physical and neutral damage excluded; some special boss attacks also bypass combat immunity. At ordinary maximum ranks, Blade Seal costs 190 MP with a 25-second base cooldown, Veil of Swords costs 384 MP with 90 seconds, and Magic Defense lasts 22 seconds with 120 seconds. Bonus ranks change values. Cooldown reduction applies only where the skill permits it: Magic Defense and Protective Aura have restricted cooldown reduction and compete with each other. These windows are not a guarantee of permanent invulnerability.

Choosing a character

Choose the role and handling you enjoy, then read the conditions beside the tier. A support specialist can be the right choice for a struggling party even with lower personal damage. An open tier is a request for better comparable evidence, not a reason to exclude that character from a party.

Chaos Bible

A detailed companion for your adventures in Chaos Lunia. Availability, event rewards, and balance can change. Read the conditions beside each rate and check the current in-game tooltip before spending valuable materials.