In　the　present　study,　the　thermal　stability　of　graphene-reinforced　composite　laminates　(GPL-RC)　with　diverse　functional　gradients　and　width　delamination　layers　is　examined.　In　this　regard,　various　models　of　laminated　GPL-RC　are　considered　with　different　geometrical　and　material　parameters.　Utilizing　the　physics-informed　neural　networks　(PINN),　we　calculate　the　energy　release　rate　(ERR)　at　the　cleavage　boundary,　aiming　to　gauge　cleavage　growth　potential.　This　study　also　delves　effects　of　various　graphene　reinforcement　distributions　and　delamination　configurations　on　the　vibrational　attributes　of　delaminated　GPL-RC　sheets,　with　an　emphasis　on　pre/post　heat　bending　modalities.　Solutions　are　grounded　in　the　third-order　shear　strain　theory　(TSDT),　integrating　von　Karman　geometric　nonlinearity.　Using　the　principle　of　minimal　potential　energy,　the　nonlinear　equilibrium　equations　are　tackled　using　PINN.　Theoretical　insights　obtained　are　verified　via　a　comparison　to　other　published　studies.　Notably,　parametric　experiments　indicate　that　the　ERR　in　the　FGX　configuration　in　which　most　reinforcement　material　located　adjacent　to　the　upper　and　lower　surfaces　of　the　plate,　is　double　that　of　the　FGA,　in　which　most　reinforcement　material　adjacent　to　the　lower　surface　of　the　plate.　Moreover,　while　the　FGX　sheet＇s　fundamental　frequency　surpasses　other　graphene　configurations　at　the　primary　temperature,　its　natural　frequency　in　the　post-buckling　modality　is　notably　the　least　compared　to　the　entire　sample　set.　©　2025　Elsevier　Ltd