Strongest pals

Rank the whole plan, not a base Attack column

The reference’s thirteen case files combine eighteen species because pairs and party support can matter more than one attacker. This is an attributed build argument, not a measured universal ladder. Use the appended current-snapshot stats and partner ranks as inputs, then test your intended boss.

Match the candidate to the role

Four circuits to test, not guaranteed wins

• Electric: mounted Helzephyr Lux conversion, Snock buildup and Slowatt’s conditional discharge. At rank 5 the captured Slowatt value is60; that is not60% extra single-target damage in every situation.
• General raid frame: Solenne’s distinct-species player support, Orserk hit stacks, Lyleen healing and a damage flex. Rank 5 Solenne is80 in the snapshot, with stacking disabled.
• Dragon/Fire mount frame: Eidrolon Ignis gains per eligible other teammate; at rank 5 its raw per-member value is6, so four eligible companions model24. Verify dual-element counting and conditions.
• Dark frame: conversion support plus Celesdir Noct, Solenne and a damage dealer. Health-for-damage and same-species constraints must remain visible.

Invest after the first controlled clear

Start with an attainable core and fill one missing role. Record boss, team, partner ranks, passives and player weapon. Repeated clear time and resource use are stronger evidence than a screenshot of the highest buff stack. The four circuits and the source order have not been independently benchmarked on v1.0.5.

Current-snapshot inputs for the case studies

These are species base parameters, not level-scaled final stats. Their order groups the discussed roles and does not assert a new performance ranking.