Smart Topology (meshy-t2)
Cleaner topology, separated parts, and a face budget you can set in the Image to 3D API.
Meshy's new meshy-t2 model, under the Smart Topology mode in the Image to 3D API, focuses on what makes generated geometry usable after the demo: cleaner structure, natively separated parts, triangle output, and a target face count you set with target_polycount. For game teams, explicit geometry budgets and separate parts matter more than another fidelity bump. They make an asset easier to select, revise, rig, replace, and optimize. It's API-only for now (the older lowpoly mode is deprecated but keeps serving for backward compatibility), and edge flow, UVs, and deformation still deserve a hands-on look before you call any result animation-ready.
Meshy AI
Best for quickly blocking out scenes or filling them with 3D assets without modeling from scratch.