The Strongest 3D Printing Material for Functional Parts

Before any numbers are useful, work out which of these describes how your part will fail. Most people can answer it in one sentence once they are asked.

• Tensile strength. How much pull it takes to break the material, measured in megapascals to ISO 527 or ASTM D638. Relevant to brackets under steady load, hooks, and anything that carries weight.
• Impact toughness. How much energy it absorbs before it shatters, measured in kJ/m2 by notched Charpy to ISO 179 or in J/m by Izod to ASTM D256. Relevant to anything that gets dropped, knocked or clipped on and off.
• Layer adhesion. Tensile strength measured in the Z direction, across the layer boundaries. This is the number that decides most real failures and the one least often quoted.
• Heat resistance. Heat deflection temperature. A part that is perfectly strong at 20°C and useless at 70°C has not really failed on strength at all.

Stiffness is a fifth and separate property. PLA is one of the stiffest common filaments and one of the least tough, which is exactly why it feels solid right up to the moment it snaps.

The chart below uses Bambu Lab data exclusively, because a single manufacturer testing to one standard on one specimen geometry is the only way to compare fairly. The pattern it shows is the single most useful thing in this article.

Read the gap between the two bars, not the height of the first one. PLA Basic goes from 35 to 31 MPa, an 11 percent drop. PA6-CF goes from 102 to 48 MPa, a 53 percent drop. PET-CF loses the same 53 percent.

The mechanism is straightforward. Chopped carbon fibre aligns with the extrusion path during printing, so it reinforces along the layer and does nothing across it. You buy enormous in-plane strength and inherit an equally enormous weak axis. A carbon-fibre nylon part loaded across its layers is only about 55 percent stronger than a PLA part in the same orientation (48 MPa against 31 MPa), despite costing roughly three times as much per kilogram ($60–90 against $19–30).

This is why orientation instructions matter more than material selection on most jobs. Tell your maker which direction the load comes from and let them orient the part so the load runs along layers rather than across them. That change is free and routinely worth more than an upgrade from PLA to nylon.

Switch the question from "what breaks under a pull" to "what shatters when dropped" and the ranking changes. This is where a lot of material advice goes wrong, because a stiff high-tensile material can be the worst possible choice for a part that gets handled.

Two things worth pulling out of the impact data. First, ABS and polycarbonate remain excellent toughness choices despite modest tensile numbers: Polymaker publishes 18.0 kJ/m2 for PolyLite ABS and 21.3 kJ/m2 for PolyMax PC, both well above any unfilled PLA or PETG. If your part gets dropped, that matters more than tensile strength does.

Second, Prusa does not publish a notched Charpy figure for Prusament PLA at all. Their data sheet records it as not applicable, which is a polite way of saying the specimens do not produce a meaningful result. Treat any source quoting a precise notched impact figure for PLA with suspicion.

Numbers from three manufacturers. They are grouped by vendor deliberately: comparing down a column across vendors is unreliable, comparing within a vendor is not.

Look at the Prusament rows again. Every one of them shows the vertical specimen equal to or stronger than the horizontal one, which would mean layer adhesion is better than in-layer strength. That is not a real material property; it is an artefact of specimen geometry and how yield is defined. Prusa reports tensile yield strength and Bambu Lab and Polymaker report tensile strength at break, and those are different measurements on the same curve.

The practical rules that follow:

• Compare within one manufacturer, never across.
• Check whether a number is yield or ultimate before quoting it.
• Check the print direction. An unlabelled tensile figure is almost always the flattering in-plane one.
• Check the conditioning. Nylon figures are usually published dry, and dry is not how your part will live.

If you want a simpler entry point into the two materials most jobs actually choose between, our PLA vs PETG filament guide covers that comparison without the datasheet archaeology.

Carbon-fibre nylon tops the tensile chart, and it comes with the largest caveat in the whole category. Polymaker publishes both dry and conditioned figures for Fiberon PA6-CF20: 109.3 MPa dry, 54.7 MPa saturated in the X-Y plane. Notched impact moves the other way, from 11.0 kJ/m2 dry to 35.6 kJ/m2 saturated.

Read the conditioning before you design against those numbers. Polymaker's "wet" specimens were annealed, then immersed in 60°C water for 48 hours to 5.3 percent moisture content. That is a deliberate worst case; the same data sheet puts equilibrium absorption in air at about 3.3 percent, so a part living in a humid workshop lands somewhere between the two columns rather than at the bottom of them. The direction is what matters: nylon takes up water, and as it does the part gets weaker, tougher and slightly larger. None of that is a defect; it is what polyamide does. But a nylon part specified against a dry data sheet and never dried again is not operating at its dry numbers.

If you order a nylon part, ask three questions: was the filament dried before printing, was the finished part dried and sealed for shipping, and will it be used somewhere humid. If the answer to the last one is yes, design against the saturated numbers.

Before paying a nylon premium, exhaust the free options. Most printed parts that break do so because of how they were printed, not what they were printed from.

• Orientation. Rotate the part so the principal load runs along layers. On a carbon-fibre nylon part this is worth up to a 2x difference on its own.
• Perimeters, not infill. Four or five perimeters at 20 percent infill is stronger in bending than two perimeters at 60 percent, and uses less material.
• Fillets. Sharp internal corners concentrate stress and are where printed parts crack. A 2mm fillet costs nothing.
• Wall thickness at load points. Screw bosses and clip arms want local thickening, not a global infill increase.

Our 3D printing cost guide explains how these choices feed into what a job is quoted at, which matters because perimeters and material class both push the price in the same direction.