389 independently verified
benchmark cases.
Across 18 of Enginus's 19 calculators, checked against published code tables, hand-derived first-principles calculations, real named worked examples, and independent audit-verified benchmarks — not just "the code runs without crashing."
Last audited 2026-10-18. This is a live count of current test coverage, not a certification — see "What this isn't" below.
What counts as “verified” here
Every calculator on Enginus that has its own calculation engine has a benchmark test suite that runs its actual underlying formulas against known values before that code ships (Design Basis Report doesn't — it assembles data from other calculators' already-validated engines rather than computing anything of its own, so there's no separate suite to have). We went through every suite that does exist — every file, every individual check — and separated genuine independent verification from internal software testing that doesn't belong in a claim like this one.
A published code table
EN 1992-1-1 Table 3.1, EN 1997-1 Annex D, EN 14080:2013, ACI 318 tables, ASCE 7's published load-combination coefficients, EC6's masonry formulas, and others — the calculator's output is checked directly against the number the standard itself publishes.
A hand-derived first-principles calculation
Worked out independently — by hand or in a standalone script — from the governing formula, without calling the calculator's own code, then hardcoded as the expected value.
A real, named worked example
Classic textbook problems with known published answers — like the Hibbeler/Kassimali Pratt truss below — reproduced through the actual calculator engine.
An independent audit-verified benchmark
A specific worked example whose result was independently confirmed during a dedicated code-quality audit of that calculator, separate from its everyday test suite.
A sample, not the full list
15 representative cases pulled directly from the real benchmark files, values exactly as computed. The full set lives in each calculator's own __tests__/*.bench.ts file in the codebase.
| Calculator | What was checked | Reference | Result |
|---|---|---|---|
| Material Properties | Concrete C30/37 — mean tensile strength (fctm) | EN 1992-1-1 Table 3.1 | ≈ 2.9 MPa |
| Material Properties | Glulam GL28h — fm / ft,0 / fc,0 / fv | EN 14080:2013 Table 5 | 28 / 22.3 / 28 / 3.5 MPa |
| Material Properties | Masonry fk — Clay Group 1, general-purpose mortar, fb=10, fm=5 | EN 1996-1-1 §3.6.1.2 (hand-computed) | ≈ 4.47 MPa |
| Retaining Wall Design | Bearing capacity factor Nγ at φ=30° | EN 1997-1 Annex D | 20.09 |
| Truss Analysis | 4-panel Pratt truss — diagonal force B0-T1 (joint equilibrium) | Hand-derived joint equilibrium | −250.0 kN (compression) |
| RC Slab Design | Interior panel negative-moment fraction | ACI 318-19 Table 8.10.4.1 | 0.65 × Mo |
| Load Combination Generator | ASD seismic combo 9 | ASCE 7 published coefficients | (1.0+0.105·SDS)D + 0.525ρE + 0.75L + 0.75S |
| Connection Design | Bolted double-angle shear connection, utilization | EN 1993-1-8 Table 3.4 (hand-derived) | 26.57% |
| Unit Converter | 1 inch → millimetres | International exact inch (ISO) | 25.4 mm exact |
| Rebar Calculator | EC2 anchorage length — T16, C30/37, B500, good bond | EN 1992-1-1 (independent hand calculation) | matches within tolerance |
| Wind Load Calculator | Default worked example — total wind force | Audit-verified benchmark | 937.9 kN |
| Steel Beam Design | IPE300/S355 — bending resistance Mc,Rd | Published Wpl,y = 628.4 cm3 section table | 223.08 kN·m |
| Foundation Design | Minimum flexural reinforcement, Grade 60 (413.7 MPa) | ACI 318-19 §7.6.1.1 | ρmin = 0.0020 |
| Column Interaction Diagram | EC2 rectangular column — balanced-point moment | Strain-compatibility hand derivation | matches within tolerance |
| RC Column Design | EC2 slenderness limit λlim, 300×300 column, φef=1.5, single-curvature end moments | EN 1992-1-1 §5.8.3.1 (hand-derived) | λlim ≈ 35.10 |
Coverage varies by calculator
Some calculators have a long, mature benchmark history; others are newer or don't yet have an independent verification suite of their own. Shown honestly, including the zeros — and Design Basis Report, marked N/A rather than 0, since it has no calculation engine of its own to verify in the first place.
Not an independent calculation engine — it assembles data from already-validated calculators. See Material Properties and Load Combination Generator for the underlying verified engines.
Truss Analysis and Frame Analysis also share a stiffness-method math engine (steelAxial.ts for member buckling/tension capacity, matrix.ts for the linear solve) with 18 more independently-verified cases underneath both tools, not counted in either calculator's own total above.
What this isn't
As always: Enginus supports engineering calculation and preliminary design workflows. Final engineering decisions, verification and professional responsibility remain with the project engineer.