Benchmarks

Measured, not asserted.

We evaluate Sietch on ProgramBench — reconstruct source from behavior alone, then check that it passes the original program's behavioral tests.

93%
Measured without source access, disassembly, or internet.

Functional-equivalence success reconstructing source on the easy & medium ProgramBench split.

Easy99%
Medium88%
HardIn testing
How we measured this

Behavior in, source out

Each target is evaluated binary-in, source-out — no source access, no decompilation, no internet.

Pass = behavioral match

A pass means every behavioral test for that target passes against our reconstruction — not a diff, an equivalence.

The set

ProgramBench: 200 programs, 248,000+ behavioral tests, graded easy / medium / hard.

Full detail

Complete methodology and per-target results are available under diligence — get in touch.

More soon

Architecture, per-target results, and deeper technical detail — on the way.