Unverified50% confidenceFactExact time
UltraShape 1.0采用端到端的单阶段生成架构，直接从输入图像预测完整的3D资产，省去中间的多视图生成步骤
1
Sources
50%
Confidence
Long-term
Relevance
7/13/2026
First Seen
Sources
开源3D生成模型实测：Trellis 2能否挑战付费产品
Related Claims
Unverified腾讯混元3D系列采用多阶段生成范式：先通过图像生成多视图，再经由重建网络得到网格，最后进行纹理烘焙74% similarUnverifiedThe breakthrough of generative AI in the 2D image domain is extending toward 3D content generation, with mainstream technical paths including NeRF-based implicit 3D representation, diffusion model-based point cloud generation, and methods that reconstruct 3D meshes from 2D multi-view images.71% similarUnverifiedSenseTime is reportedly developing a multimodal image large model called U1 Pro, internally benchmarked against GPT Image 2, targeting design scenarios, supporting long-range generative review loops, and capable of outputting at up to 8K resolution.70% similarUnverified3D资产通常为OBJ、FBX或glTF格式的三维模型，在AI产品图生成中作为真实性锚点，系统先从3D模型渲染特定角度视图作为条件输入68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/503054API
curl https://kongchang.com/api/v1/knowledge/claims/503054MCP
get_claim(id=503054)