New Models to Know: MoE-ViE, τ0-VLA, 4DAnyone, and More
TheTuringPost · x · 2026-08-26
Notable new models this week include:
- MoE-ViE: Meta's sparse mixture-of-experts vision encoder for efficient image and video understanding.
- τ0-VLA: A robot foundation model that "thinks" through possible outcomes before acting.
- 4DAnyone: Turns a single casual video into an animatable 4D human (Ant Group & Zhejiang University).
- WithEveryone: Tencent's image model for generating group scenes while preserving identity.
- PixRestore: A compact diffusion transformer for pixel-space image restoration (OPPO & PolyU).
Related event: This Week's New Models: Meta Vision Encoder and Robot Foundation Models(2 posts)→
More from Models
- Grok Voice 3.0 reportedly powers Tesla in-car assistant with millions of daily conversations — XFreeze · 2026-10-15
- Gemini 4 Argon coding style: neat one-line comments, but does the code actually work? — zacharynado · 2026-10-15
- Why ChatGPT beats the better Claude with mainstream users: being first known — aitrendz_xyz · 2026-10-15
- 'You need high pain tolerance to review Gemini 4,' researcher quips amid error screens — gleech · 2026-10-15
- MiniCPM-V 4.7 runs a 35B MoE multimodal model on phones with just 3GB memory — nevrekaraishwa2 · 2026-10-15
- Gemini 4 Argon wrote 5,000+ lines of pure Python to build its own ray-tracing 3D engine — aitrendz_xyz · 2026-10-15