Logical Intelligence Tops Leading AI Verification Benchmarks as Verified Code Generation Nears Reality with Aleph

▤ Summary

Aleph achieved top scores on four formal reasoning benchmarks—PutnamBench, VeriSoftBench, LeanEval, and Verina—solving 99.4% of Putnam problems, 94% on VeriSoftBench, and 100% on Verina. These results demonstrate that machine-checkable, formally verified code generation is practical for mission-critical systems rather than only a theoretical aim. Aleph is already running in limited production pilots, including verification work on the Ethereum Foundation's ArkLib for zkEVM infrastructure, and Logical Intelligence plans a public beta later this year. The company positions Aleph as an end-to-end formal verification agent that integrates into engineering workflows to replace slow manual verification and flag failure modes before deployment.

▦ Linked Companies