Infill is the internal structure that gives FDM printed parts their bulk strength, and the choice of infill pattern and density has more impact on part performance per unit of material than almost any other slicer setting. The Prusa Research infill patterns documentation catalogs the core options available in PrusaSlicer and summarizes the key tradeoffs: 2D patterns like grid and lines are fast to print but directional in their strength; 3D patterns like gyroid and honeycomb distribute strength more isotropically at the cost of slower printing and more material; and structural patterns like lightning sacrifice strength entirely for maximum speed and material savings when internal structure is only needed for top surface support. Understanding which scenario calls for each pattern — and what density percentage achieves the practical strength target for a given application — eliminates the common mistake of defaulting to 20% grid for every print regardless of whether the part needs to bear any load.

Grid and Lines: Fast, Directional, Practical

Grid infill prints in two perpendicular directions on alternating layers, creating a rectilinear lattice that is strong along both primary axes but weaker under diagonal loading. Lines infill (Rectilinear in some slicers) prints in a single direction per layer, alternating 90° between layers — faster than grid by avoiding direction-reversal moves at each intersection, and producing slightly better bonding because continuous paths reduce seam count. At 20% density, grid or lines provides adequate support for top surfaces and moderate structural loads. At 40%, the structure becomes noticeably stiffer. Grid and lines are the correct default for functional prints where loading direction is known and consistent — brackets, mounts, boxes, and frames benefit from the predictability and speed of simple rectilinear infill over more complex patterns that provide limited advantage when loading direction is predetermined.

Gyroid: The Engineering Default for Isotropic Parts

Gyroid is a triply periodic minimal surface — a mathematically continuous structure that extends through 3D space without self-intersection, producing a print path that never requires the toolhead to reverse direction mid-layer, which enables higher print speeds without quality loss. More importantly for functional parts, gyroid infill distributes stress in all three axes approximately equally, producing parts with similar strength regardless of which direction the load is applied. This isotropy makes gyroid the recommended default for parts whose loading direction is uncertain or variable: phone cases, general-purpose enclosures, tool handles, and structural components where impact loads can come from any direction. Gyroid also produces excellent support for top surface layers because the curved paths eliminate the flat gaps between infill lines that cause top surface dimpling at low infill densities. At 15% gyroid, most prints have fully supported top surfaces with clean layer bonding. At 25%, the part is noticeably rigid and handles moderate mechanical loading. Print time compared to grid at the same density is similar or slightly longer due to the 3D path complexity, but the strength per unit of material is measurably higher under off-axis loading.

Honeycomb and Cubic: Structural Alternatives

Honeycomb infill generates a hexagonal cell structure that is highly efficient at distributing compressive loads — a pattern borrowed from nature's own structural engineering. It performs well under uniform compression (parts that sit on flat surfaces under vertical loads) but less efficiently under tension or shear than gyroid. The print speed penalty is noticeable because the hexagonal corners require short acceleration-deceleration sequences that slow the overall path. Cubic infill (called 3D Honeycomb in some slicers) extends the hexagonal logic into three dimensions, producing a truncated octahedron cell structure that improves off-axis strength over 2D honeycomb. Both patterns are good choices for load-bearing printed parts like furniture components, machine frames, and anything that will experience primarily compressive loading from a predictable direction. Gyroid at equivalent density outperforms honeycomb under most real-world mixed loading scenarios, which is why gyroid has largely displaced honeycomb as the recommended structural infill in current community practice — but honeycomb remains relevant for specifically compressive applications where its efficient cell geometry is well-matched to the load case.

Lightning Infill: Speed Over Strength

Lightning infill generates the minimum internal structure required to support the top surface of the print — a branching, tree-like network that reaches just enough of the top surface area to prevent visible dimpling, with no additional material. It is the fastest infill option by a significant margin (typically 30–60% material reduction versus 20% grid) and is the correct choice for any part where internal strength is irrelevant: figurines, decorative objects, display models, prototypes that will only be visually evaluated, and any print that will be solid on the outside but hollow functionally. Lightning at typical settings provides no meaningful contribution to part strength under mechanical loading — bending or crushing a lightning-infilled part produces easy deformation and fracture through the minimal internal structure. The use case is speed and material efficiency where the print exists for appearance or form rather than function, and lightning excels at that use case without compromise.

Density, Top Layers, and Surface Quality

Infill density interacts with top surface quality through the support it provides to the top skin layers. At very low density (under 10%), the top skin bridges across long gaps between infill contact points, producing visible dimpling and rough texture. Increasing to 15–20% typically resolves top surface dimpling on standard geometries. For the smoothest possible surface, three to four top layers at 100% infill combined with ironing — available in Cura and PrusaSlicer, it passes the nozzle across the top at low flow rate to reflow and flatten the surface — produces glass-smooth horizontal results on well-tuned machines. Ironing adds 10–20% print time but the quality improvement on flat cosmetic surfaces is dramatic. The combination of 20–25% gyroid with 4 top layers and ironing is the recommended profile for display-quality functional parts.

What It Means for Makers

The practical recommendation: 15–20% gyroid for general functional prints, grid or lines at 20–30% when speed matters and loading direction is known, lightning for decorative parts. Resist the instinct to increase infill density as a response to part failure — printed parts almost always fail at layer interfaces or geometric stress concentrations, not from insufficient infill. Adding perimeter walls (increasing wall count from 2 to 3 or 4) increases strength more effectively than increasing infill density, because walls carry bending and torsional loads more efficiently than internal fill. Optimize perimeter count first, then infill density.

Sources