Quick answer: Use 10-15% infill for decorative prints, 15-20% for general-purpose functional parts, and 25-40% for anything under real structural load — going past 30-40% rarely helps, since Prusa's own knowledge base says higher density "isn't commonly necessary" once wall count is dialed in. For pattern, gyroid gives the best strength-to-material ratio (Ultimaker found 20% gyroid ≈ 25% grid in strength) at a 5-15% print-time cost versus grid at the same density.

Most infill advice stops at “20% is a good default” without explaining when that number is wrong, or that the pattern you pick matters almost as much as the percentage. Infill controls three things at once — strength, print time, and material cost — and the biggest mistake isn’t picking the wrong percentage, it’s reaching for infill when the real fix is adding a wall.

Infill by the numbers

Infill percentage by use case

Use caseInfill %Why
Miniatures, display models, decorative prints0-10%No structural load — just enough for clean top-layer bridging
General-purpose functional prints (brackets, organizers)15-20%Prusa's own default profile range; sweet spot for most prints
Structural parts, light mechanical use25-40%Gyroid at 30% feels nearly solid under hand pressure
End-use parts under repeated load or stress40-60%Pair with 3+ perimeters — walls matter more than infill here

Pattern comparison: strength, speed, and when to use each

PatternStrength profilePrint speedBest for
GyroidNear-isotropic — equal strength from every direction5-15% slower than gridFunctional parts, anything under load from an unknown direction
GridStrong on X/Y axes, weak at 45°Fast, common defaultGeneral-purpose prints, quick prototypes
CubicEven strength in 3D, good for large printsSimilar to gridLarge parts where weight reduction matters
HoneycombVery strong per gram, but slow to printNoticeably slower than gyroid or gridMax strength-to-weight where print time isn't a priority
LightningMinimal — supports top layers onlyFastest, least materialPrototypes, draft prints, anything not load-bearing

1. Percentage matters less than most guides claim

Infill percentage gets treated as the single dial that controls part strength, but Prusa’s own knowledge base is blunt about the ceiling: going past 30% infill “isn’t commonly necessary,” and if a part still feels weak at that density, the fix is adding perimeters (wall loops), not cranking infill further. Walls carry impact resistance and much of the load path in a typical FDM part; infill mainly keeps top and bottom layers from sagging into open space and adds some compressive resistance. A part printed at 15% infill with 4 perimeters will often out-perform the same part at 50% infill with 2 perimeters, at a fraction of the print time and filament.

Check filament spools on Amazon →

Stocking filament and parts for a shop? A free Amazon Business account unlocks quantity discounts and tax-exempt purchasing on the extra spools that higher-infill test prints burn through.

2. Gyroid is the default worth switching to

Grid ships as the default pattern in most slicers because it’s fast and predictable, but it has a real weakness: strong along its own grid lines, noticeably weaker at 45° to them. Gyroid’s continuous, curved internal geometry has no preferred direction, so it resists force about equally no matter which way a part gets loaded — which is why Ultimaker’s own comparison found 20% gyroid infill performs roughly on par with 25% grid infill. The cost is print time: gyroid’s more complex toolpath runs 5-15% slower than grid at the same density. For anything functional where the load direction isn’t fully predictable — brackets, mounts, replacement parts — that trade-off is almost always worth it.

3. Pick a pattern by what the part actually does

Not every print needs gyroid. Cubic infill spreads strength evenly through a part in three dimensions and is a reasonable pick for large prints where cutting weight matters. Honeycomb delivers strong per-gram numbers but is one of the slower patterns to print, so it’s a fit when strength-to-weight is the priority and time isn’t. Lightning infill only builds support directly under features that need it instead of a repeating lattice through the whole part, making it the fastest and most material-efficient option for prototypes, drafts, or anything that isn’t load-bearing at all. Miniatures and display prints often skip infill questions entirely — 0-10% is standard, since there’s no structural demand and the goal is just clean top-layer bridging.

4. Setting it in your slicer

Infill density and pattern live in the same settings panel in every major slicer (PrusaSlicer, Cura, OrcaSlicer, Bambu Studio). Start from the use-case table above, set pattern to gyroid for anything functional, and print a single test part before committing a full batch — infill changes affect print time and material estimates immediately, so the slicer’s own time/weight estimate is a fast sanity check before a long print starts. If a finished part flexes or cracks under normal use, add a perimeter before reaching for more infill; if it’s only sagging on top surfaces, increase top layer count before touching infill at all.

Check calibration filament on Amazon →

Infill, in one paragraph

Match infill percentage to the part’s actual job: 0-10% for decorative prints, 15-20% for general functional use, 25-40% for structural parts, and 40-60% only for end-use parts under repeated load — paired with more perimeters, not instead of them. Gyroid is the pattern worth defaulting to for anything functional since it resists force from every direction and beats grid’s strength at a lower density, at a modest 5-15% print-time cost. Past 30-40% infill, more density stops paying off; a part that still feels weak needs another wall loop, not another infill percentage point.