This release is centered on strengthening Ziren’s path toward proving larger Geth- and keeper-oriented workloads.
Ziren already patches the Go runtime’s GC and scheduler, and has reached roughly 4x instruction-overhead efficiency relative to Reth. This makes it possible to move toward real-time proving for Ethereum’s most widely used execution client. Alongside this, the release hardens the Linux ABI paths these workloads depend on, improving correctness at the boundary between guest execution and the proving system.
In parallel, the release includes syscall AIR hardening, malformed-input verifier protections, core-to-precompile lookups for arguments and results, and broader correctness fixes across arithmetic, curves, bitwise logic, and division handling.
This is about strengthening the parts of the stack developers ultimately have to trust.
v1.2.5 expands shape generation for larger Geth-oriented proving workloads through new and regenerated shapes, a fix for build-shapes out-of-memory issues, and support for shapes covering 100 Geth blocks.
This improves the infrastructure around larger proving configurations and makes heavier execution-client-style workloads easier to work with in practice.
The larger story in v1.2.5 is hardening around real execution paths.
The release includes major syscall AIR work, Linux-syscall-related fixes such as the is_linux_syscall unconstrained bug and a patched Go runtime time syscall path, verifier protections for malformed inputs, core-to-precompile lookups for arguments and results, and multiple correctness fixes across core components including arithmetic, bitwise logic, curve parameters, shift behavior, and range handling.
The result is reduced fragility across the proving stack and tighter correctness in places that affect real execution and verification.
The release also keeps the Groth16 verification key unaffected by Ziren upgrades, while adding historical VK bundling and public-input verification support.
This reduces verification-side disruption across upgrades and makes the system easier to maintain. It also matters for BitVM2/3-style systems, where keeping the proof system upgradeable is imperative. In practice, this supports systems such as GOAT Network’s BitVM-based architecture, where proving-system maintainability is key for long-term usability.
In short, v1.2.5 improves correctness, stability, and integration reliability across the prover stack. It strengthens the path toward larger Geth- and keeper-oriented workloads, hardens Linux ABI and syscall-related execution paths, and provides a more stable upgrade path for integrators, including BitVM2/3 systems.
If you’re building systems that need a production-oriented zkVM with stronger execution correctness, better support for larger proving workloads, and lower integration friction, start building with Ziren today.
Read the full release: https://github.com/ProjectZKM/Ziren/releases/tag/v1.2.5
This release is centered on strengthening Ziren’s path toward proving larger Geth- and keeper-oriented workloads.
Ziren already patches the Go runtime’s GC and scheduler, and has reached roughly 4x instruction-overhead efficiency relative to Reth. This makes it possible to move toward real-time proving for Ethereum’s most widely used execution client. Alongside this, the release hardens the Linux ABI paths these workloads depend on, improving correctness at the boundary between guest execution and the proving system.
In parallel, the release includes syscall AIR hardening, malformed-input verifier protections, core-to-precompile lookups for arguments and results, and broader correctness fixes across arithmetic, curves, bitwise logic, and division handling.
This is about strengthening the parts of the stack developers ultimately have to trust.
v1.2.5 expands shape generation for larger Geth-oriented proving workloads through new and regenerated shapes, a fix for build-shapes out-of-memory issues, and support for shapes covering 100 Geth blocks.
This improves the infrastructure around larger proving configurations and makes heavier execution-client-style workloads easier to work with in practice.
The larger story in v1.2.5 is hardening around real execution paths.
The release includes major syscall AIR work, Linux-syscall-related fixes such as the is_linux_syscall unconstrained bug and a patched Go runtime time syscall path, verifier protections for malformed inputs, core-to-precompile lookups for arguments and results, and multiple correctness fixes across core components including arithmetic, bitwise logic, curve parameters, shift behavior, and range handling.
The result is reduced fragility across the proving stack and tighter correctness in places that affect real execution and verification.
The release also keeps the Groth16 verification key unaffected by Ziren upgrades, while adding historical VK bundling and public-input verification support.
This reduces verification-side disruption across upgrades and makes the system easier to maintain. It also matters for BitVM2/3-style systems, where keeping the proof system upgradeable is imperative. In practice, this supports systems such as GOAT Network’s BitVM-based architecture, where proving-system maintainability is key for long-term usability.
In short, v1.2.5 improves correctness, stability, and integration reliability across the prover stack. It strengthens the path toward larger Geth- and keeper-oriented workloads, hardens Linux ABI and syscall-related execution paths, and provides a more stable upgrade path for integrators, including BitVM2/3 systems.
If you’re building systems that need a production-oriented zkVM with stronger execution correctness, better support for larger proving workloads, and lower integration friction, start building with Ziren today.
Read the full release: https://github.com/ProjectZKM/Ziren/releases/tag/v1.2.5
