Rocq
A proof assistant based on the calculus of inductive constructions, with a long history of use in program verification and mathematical formalization.
Timeline (last 90 days)
Rocq is the name of the Coq proof assistant after its renaming in 2025, and is an interactive theorem prover based on dependent type theory.
Anthropic's proof pipeline is: compile open-source libraries with Faust tests into RISC-V executables, layer on the RISC-V and operating-system formal semantics of FreeRTOS plus the non-crash property, mechanically generate specifications, have Claude write proofs, and hand them to the Rocq kernel for checking.
These proofs total roughly 180 million lines of Rocq, with a typical proof at about 600,000 lines and a median binary of about 18,000 RISC-V instructions.
The OjasX repository contains 558 machine-verified Lean 4 theorems, 28 Rocq 9.2 proof files, and 840 automated PyTorch tests.
The authors constructed an end-to-end benchmark covering four formal languages, including Dafny contracts, Lean 4, Isabelle, and Rocq theorem statements, totaling 3,234 lines of paired data.
Rocq (formerly Coq) is a proof assistant based on the calculus of inductive constructions, with long-standing accumulation in the fields of program verification and mathematical formalization.
All Facts (6)
Rocq is the name of the Coq proof assistant after its renaming in 2025, and is an interactive theorem prover based on dependent type theory.
50%UnverifiedAnthropic's proof pipeline is: compile open-source libraries with Faust tests into RISC-V executables, layer on the RISC-V and operating-system formal semantics of FreeRTOS plus the non-crash property, mechanically generate specifications, have Claude write proofs, and hand them to the Rocq kernel for checking.
50%UnverifiedThese proofs total roughly 180 million lines of Rocq, with a typical proof at about 600,000 lines and a median binary of about 18,000 RISC-V instructions.
50%UnverifiedThe OjasX repository contains 558 machine-verified Lean 4 theorems, 28 Rocq 9.2 proof files, and 840 automated PyTorch tests.
50%UnverifiedThe authors constructed an end-to-end benchmark covering four formal languages, including Dafny contracts, Lean 4, Isabelle, and Rocq theorem statements, totaling 3,234 lines of paired data.
50%UnverifiedRocq (formerly Coq) is a proof assistant based on the calculus of inductive constructions, with long-standing accumulation in the fields of program verification and mathematical formalization.
50%