What Infill Actually Does
Infill is the internal lattice structure printed inside a model, between the outer walls (perimeters) and the top/bottom solid layers. Almost nothing you print is solid plastic all the way through — even at "100% infill," the slicer is still printing a repeating pattern, just packed so tightly there's no visible gap.
Infill exists for three reasons: it gives the top layers something to print on top of (without it, top surfaces would sag into open air), it adds mechanical strength and stiffness to the part, and it controls how much filament and time the print consumes. Every infill decision is a trade-off between those three things.
The pattern you choose changes how that trade-off plays out. Two prints at the same 20% infill density can have noticeably different strength, weight, and print time depending purely on which pattern fills that 20%.
Infill percentage has a far bigger effect on strength and print time than pattern choice does. If you're only going to change one setting, change the percentage first. Pattern choice is the fine-tuning step after that.
Infill Percentage: The Bigger Lever
Infill percentage controls how much of the model's interior volume is filled by the pattern — 0% is hollow (just walls and top/bottom layers), 100% is fully solid. Most slicers default to 15-20%, which is enough for the majority of prints.
The relationship between infill percentage and strength is not linear. Going from 10% to 20% infill makes a noticeable difference in stiffness and impact resistance. Going from 50% to 60% makes almost no difference — you're adding 10% more material for a fraction of a percent more strength. This is why "more infill = stronger" is true but massively diminishing-returns past about 30-40%.
The Patterns, One by One
Grid
Two sets of parallel lines crossing at 90°, forming a square lattice when viewed from above. It's fast to slice and prints quickly because the lines are long and continuous. Grid is strong in the X-Y plane it's printed in but has almost no structure resisting force along the Z axis — a grid-infilled part can be surprisingly easy to crush from above. Good default for flat panels and prints loaded from the side, not from above.
Gyroid
A continuous, wavy 3D curve that loops through the part in all three dimensions without ever forming a flat repeating layer. Because the structure curves through Z as well as X-Y, gyroid is roughly isotropic — it resists force from any direction reasonably evenly. It also prints with smooth, continuous toolpaths, which makes it quieter and often comparable in speed to grid despite the more complex geometry. Gyroid has become the default recommendation for general-purpose functional parts for good reason.
Honeycomb
Hexagonal cells, the classic "strong for its weight" shape borrowed from cardboard and aerospace panels. Honeycomb is genuinely strong per gram of material, but it's slow to print — every hexagon edge is a separate short segment requiring acceleration and deceleration, so print time climbs faster than with grid or gyroid at the same density. Best reserved for parts where weight is critical and print time is not, like RC aircraft components.
Cubic / Cubic Subdivision
Cubic infill prints a 3D grid of cubes, with each cube's faces oriented so forces are distributed along all three axes evenly — similar goal to gyroid but with flat faces instead of curves. Cubic subdivision additionally varies the cube size by depth, using larger cubes near the center of thick parts and smaller ones near the surface, saving material in the interior where it matters least. Good alternative to gyroid for parts that need omnidirectional strength.
Triangles
A triangular lattice — very rigid in-plane because triangles can't deform without changing the length of their sides (unlike squares, which can shear into rhombuses). Triangle infill resists in-plane shearing and torsion better than grid at the same density. Less commonly needed than gyroid or cubic, but useful for parts subject to twisting forces.
Lines / Concentric
Lines is the simplest and fastest pattern — straight parallel lines in a single direction, alternating direction each layer. Concentric follows the outline of the part, creating rings that echo the perimeter shape. Both are fast but provide minimal structural benefit beyond supporting top layers. Concentric is occasionally useful for parts where the infill pattern is visible through translucent or low-infill prints and you want it to look intentional.
Lightning
A tree-like, branching pattern that only builds support where it's structurally needed — directly underneath surfaces that need backing. Lightning infill uses dramatically less material than any other pattern (sometimes 80-90% less at the same nominal "density") because it isn't trying to fill volume uniformly, just support what's above it. It provides essentially no mechanical strength to the part itself — purely a print-time and material-saving option for non-functional prints.
| Pattern | Strength profile | Print speed | Best for |
|---|---|---|---|
| Grid | Strong in-plane, weak vertically | Fast | Flat panels, side-loaded parts |
| Gyroid | Even in all directions | Fast–medium | General-purpose functional parts (best default) |
| Honeycomb | High strength-to-weight | Slow | Lightweight strong parts, RC frames |
| Cubic / subdivision | Even in all directions | Medium | Functional parts, alternative to gyroid |
| Triangles | Resists shear/torsion | Medium | Parts under twisting load |
| Lines / concentric | Minimal | Fastest | Prototypes, visible/translucent infill |
| Lightning | Minimal | Very fast, least material | Decorative prints, draft prints |
Which Pattern for Which Print
Rather than memorizing every pattern's properties, most prints fall into one of a few categories. Here's the shortcut version:
- Miniatures, figurines, vases: lightning or gyroid at 5-10%. Vase mode prints (single perimeter, no top layers) use no infill at all.
- Everyday prints — phone stands, organizers, brackets: gyroid at 15-20%. This is the right default for the vast majority of what people print.
- Functional parts under moderate load — tool holders, enclosures with mounting screws: gyroid or cubic at 20-30%.
- Mechanical parts — gears, jigs, load-bearing brackets: gyroid or cubic at 30-50%, combined with 4-6 wall loops.
- Lightweight high-strength — drone frames, RC parts: honeycomb at 15-25%, prioritizing strength-to-weight over print time.
- Parts loaded primarily from one direction — flat plates, shelf brackets: grid, oriented so the lattice plane aligns with the load direction.
Walls vs Infill: What Actually Matters More
This is the part most guides skip. For most functional parts, wall loop count (the number of solid perimeter rings around the outside of the part) contributes more to real-world strength than infill does — especially for impact resistance and bending stiffness.
A part with 2 walls and 50% infill is often weaker than the same part with 5 walls and 15% infill, despite using less total material in the second case. This is because the outer walls form a continuous shell that resists bending, while infill mostly resists local crushing and supports the top surface.
Before cranking infill to 50%+ for a "stronger" part, try increasing wall loops to 4-6 first (in OrcaSlicer: Strength → Walls → Wall loops). It's often a bigger strength gain for less added material and print time than the equivalent infill increase.
- Part is genuinely crushed or compressed in use (clamps, jigs, vises)
- Part has threaded inserts or heat-set screws that need surrounding material
- You're printing a small part where the time/material cost of higher infill is negligible
- Part needs to resist bending or snapping (brackets, hooks, levers)
- Part will be dropped or take impacts
- You're already at 20%+ infill and want more strength without a big time cost
If you only take one thing from this guide: switch your default infill pattern to gyroid at 15-20% and stop thinking about it for the vast majority of prints. It's isotropic, prints efficiently, and covers the overwhelming majority of use cases without any downside compared to grid or lines.
Reach for honeycomb only when strength-to-weight is the actual design constraint, cubic as a gyroid alternative for parts with flat internal geometry, and lightning when a print is purely decorative and you want to save time and filament.
And remember: if a part is genuinely failing under load, check your wall loop count before you touch infill percentage again. It's usually the faster, cheaper fix.