List your company

Automated Theorem Proving on seL4

77.6Proof Success RateStepwise (Mistral)
2.7222.1641.661.04Mar 20, 2026

Evaluation results

MethodLinks
Stepwise (Mistral)Category=Ours, Backbone=Mistral-7B2026.0377.6
Stepwise (Qwen3)Category=Ours, Backbone=Qwen3-1.7B2026.0370.4
SledgehammerCategory=Symbolic2026.0340.3
FVELCategory=Neural, Model=Mistral 7B-Instruct2026.037.8
AutoCategory=Symbolic2026.035.9
SeleneCategory=Neural, Model=GPT-4o2026.035.6