An open challenge to improve a proof bound
In a blog post published Aug. 20, 2026, the Ethereum Foundation reported that better.codes was live. The open autoresearch challenge was built by the Foundation’s Formal Verification team in collaboration with Yukon and zkSecurity, according to the post.
The Foundation’s post says participants point their own AI agents at raising the machine-checked soundness bound of koalaIRS12, a Reed–Solomon proximity problem intended to advance modern succinct non-interactive proof systems, or SNARKs.
What the benchmark checks
According to the Ethereum Foundation, the Lean kernel checks every submission, and each promoted proof raises the bound toward the fixed 128-bit target. The challenge is aimed at improving a proof of the problem’s soundness bound; the post does not report that the target has been reached.
The Foundation says nearly all production hash-based SNARKs rely on proximity gaps and correlated agreement for Reed–Solomon codes. It cites proof systems securing zkrollups and zkVMs, as well as systems central to Ethereum’s post-quantum roadmap.
The 128-bit target is not a completed guarantee
The challenge’s target should not be read as a result already achieved. The Ethereum Foundation says deployed systems target 128-bit security, but that guarantee holds in full only if the underlying conjectures are true. The launch describes an effort to raise the proven lower bound toward the target; it does not say the challenge has reached it or resolved those conjectures.