CVPR 2026 Oral | 牛津 & Meta AI 推出 VGGT-Ω：前馈 3D 重建迈入 10B 参数时代，动态场景精度升
Quick Take
Oxford and Meta AI have introduced VGGT-Ω, a groundbreaking model for high-precision 3D reconstruction of dynamic scenes, achieving a 77% improvement in camera estimation accuracy on challenging benchmarks like Sintel. This advancement marks a significant leap into the 10 billion parameter era for feedforward models.
Key Points
- VGGT-Ω enhances dynamic scene reconstruction accuracy significantly.
- Achieved a 77% improvement in camera estimation on the Sintel benchmark.
- Represents a leap into the 10 billion parameter model era.
- Developed collaboratively by Oxford and Meta AI.
- Targets advancements in computer vision applications.
Article Excerpt
From source RSS / original summaryVGGT-Ω通过一套精妙的架构改进，实现了对动态场景的高精度重建，在Sintel 等极具挑战性的基准测试中，将相机估计精度足足提升了77%。
Reader Mode unavailable (could not extract clean content).
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from WebSearch (Tavily)
See more →Anthropic releases new model, Opus 4.8 - Axios
Anthropic has launched Claude Opus 4.8, an upgrade to its AI model that enhances coding and knowledge work capabilities while maintaining the same price. Although it still trails behind the upcoming Mythos-class models, Opus 4.8 outperformed competitors in key benchmarks such as agentic coding and financial analysis.
