Filament Ships an Ecosystem-Wide Performance Beta

Filament's Head of Developer Relations, Alex Six, announced a beta release on June 26, 2026, bringing performance improvements across the entire first-party Filament package ecosystem — forms, tables, panels, widgets and charts, and assorted internals. It follows on from earlier work that reworked the tables package between Filament v3 and v4, and pushes that same performance focus out to the rest of the framework.

What actually changed

The changes are broad rather than a single headline feature, which is exactly the point — this is a pass over the parts of Filament that get exercised on every request in a real admin panel.

On the forms side, large forms render faster thanks to a refactor of Blade views into plain HTML-generation functions, cutting out template overhead on the hot path. Repeaters, builders, and repeatable entries are more aggressively memoized, as is component configuration like colors, descriptions, icons, and tooltips — data that used to get recomputed on every render now gets computed once and reused. Selects also resolve their options fewer times in some cases, avoiding redundant query or option-list work.

Tables pick up a similar set of improvements: optimized fast-paths for basic text columns and actions, various internal memoizations, and fewer redundant state-formatting passes per cell. If you're rendering a table with hundreds of rows and several columns, formatting logic that ran once per cell, per render is now able to skip that repeated work. Table reordering for BelongsToMany relationships is also faster, and query-builder internals are memoized so the same underlying query isn't rebuilt unnecessarily.

Beneath the panel layer, model-to-resource lookups, resource slugs, resource page names, and relation-manager visibility checks are all memoized now — these are the kinds of lookups that happen repeatedly per request across a panel with many resources, and caching them removes work that was previously duplicated for no reason. Navigation group sorting is faster too, which matters more the more resources and custom pages a panel registers.

Widgets get attention as well: stat chart widgets no longer import the entire Chart.js library, shrinking the JS bundle shipped to the browser, and they no longer redraw when the underlying data hasn't actually changed. There's also a fix for a memory leak caused by action-modal event listeners that weren't being cleaned up — the kind of bug that quietly degrades a long-lived browser tab in an admin panel that stays open all day.

Rounding things out: a more efficient ComponentAttributeBag, fewer redundant config fetches and button component instantiations, more efficient filesystem disk caching and file-existence checks, and notification Alpine components now serializing less data to the frontend.

Why it matters

None of these are individually dramatic, but they add up in exactly the place teams tend to feel Filament's weight: panels with a lot of resources, tables with a lot of rows and columns, and forms with nested repeaters. If you're running Filament as an internal admin tool at PhpArchitect-scale — dozens of resources, wide tables, dashboards with several stat widgets — these are the code paths that get hit on nearly every page load, so shaving redundant work out of them is where performance work pays off the most.

How to try it

To pull in the beta, update your composer.json constraint for any first-party filament/* package:

• On Filament v4: ^4.12@beta
• On Filament v5: ^5.7@beta

Then run:

composer update

The Filament team is explicit that these betas are not recommended for production. They're intended for local development and staging environments while the community puts them through their paces. If you try it, report back — whether it works fine or you hit issues — on the linked GitHub discussion so the team has real-world signal before promoting the changes to a stable release.