Tripo AI unveiled Tripo P2.0 on 21 September 2026, a 3D-native foundation model that generates production-ready assets with native quad topology, according to a press release distributed by PR Newswire.
Key points
- Tripo P2.0 produces native quad-dominant meshes up to 25,000 faces and triangle meshes up to 50,000 faces, a first for AI 3D generation according to the company.
- The model generates up to four versions per prompt with different face counts and includes a Mesh Edit feature for regional regeneration without altering the rest of the asset.
- Smart UV enables single-click unwrapping of 3D assets into 2D layouts for texturing and further editing.
- Dr Yanpei Cao, chief scientist, says the model addresses the topology wall between AI-generated 3D and real production pipelines built on quad meshes.
- The release follows a P2.0 Preview in August 2026 and comes after the company raised approximately 3 billion yuan in Series B and B+ funding earlier in September.
Native quad topology changes the pipeline
Tripo AI positions P2.0 as an upgrade from Tripo Smart Mesh P1.0, with the headline advance being native quad mesh generation. Game and film pipelines are built on quad meshes because their predictable edge flow supports subdivision, rigging and animation; until now, AI-generated 3D has arrived as triangle meshes that require a separate retopology step — a manual or semi-automated process that is slow, unstable and often destroys the detail the model produced. By generating quad-dominant meshes directly, with clean part separation and edge flow the company says resembles what an artist would lay out manually, P2.0 aims to let creators move AI output straight into production.
The model supports both triangle and quad topology, offering up to 50,000 faces for triangle output and up to 25,000 faces for quad output. With a single prompt it can produce up to four versions of the same asset at different face counts, giving creators a range of detail levels without re‑prompting. The Mesh Edit feature adds regional control: a user can select part of a mesh and regenerate only that region, preserving the rest of the asset. Smart UV then unwraps the result into a 2D layout with one click, ready for texturing. Tripo AI says the model is designed for game characters, hard‑surface objects such as vehicles, props and buildings, and scalable asset creation.
Dr Yanpei Cao on the topology wall
“Topology has been the wall between AI-generated 3D and real production: game and film pipelines are built on quad meshes, and until now AI got there only through slow, unstable retopology. P2.0 can quickly generate quad-dominant meshes natively, with clean part separation and the edge flow an artist would lay out manually. Production-ready is today’s bar; the longer-term goal for Tripo is to build models that understand, generate and interact with 3D environments,” Dr Cao said in the release. The statement frames the release as a step toward world models that go beyond static asset generation.
A 3D artist receiving a character model from P2.0 could rig and animate it in a game engine without sending it through a retopology tool, and a technical artist could adjust the polygon budget by selecting one of the four generated versions rather than decimating a single high‑poly mesh.
Funding and roadmap
Tripo AI was founded in 2023 by leading scientists in AI and computer graphics. The company raised approximately 3 billion yuan in Series B and B+ funding in September 2026, according to its own announcements. The official P2.0 launch follows a P2.0 Preview released in August 2026 and adds the multi‑version generation and Mesh Edit features that were not present in the preview. Tripo AI says its plan progresses from making shapes to comprehending their construction, modifying assets already produced, and allowing motion and interplay in 3D environments.
