List your company

Automated Theorem Proving on seL4 (val)

79.8Proof Success RateStepwise (Mistral)
1.90422.12742.3562.573Mar 20, 2026

Evaluation results

MethodLinks
Stepwise (Mistral)Category=Ours, Backbone=Mistral-7B2026.0379.8
Stepwise (Qwen3)Category=Ours, Backbone=Qwen3-1.7B2026.0373.6
SledgehammerCategory=Symbolic2026.0340.5
FVELCategory=Neural, Model=Mistral 7B-Instruct2026.038.9
SeleneCategory=Neural, Model=GPT-4o2026.036.1
AutoCategory=Symbolic2026.034.9