SenseTime's SenseNova-U1.5 hits Hugging Face: 8B encoder-free multimodal model with native 4K
liuziwei7 · x · 2026-09-11
SenseTime's SenseNova-U1.5 is now on Hugging Face: a native unified 8B-MoT multimodal model that handles understanding, reasoning, and visual generation in one model.
Notably, it drops the visual encoder and VAE entirely, scaling to native 4K resolution. Weights are publicly available for download.
More from Multimodal
- Comfy offers $50 credits for cool Comfy API demos, creator vies to claim them — PurzBeats · 2026-10-01
- Dev uses AI render projection to build Silent Hill-style Three.js kitchen with Astra — nptacek · 2026-10-01
- An absurd 4-minute AI-generated Western comedy built on literal puns — Beautiful-Dealer-943 · 2026-10-01
- Ideogram's new Edit Model lands as users try AI outfit swaps — lishali88 · 2026-10-01
- Reddit user replicates Viggle-style motion transfer and character swap locally — Useful_Ad_52 · 2026-10-01
- PKU, Tsinghua, Alibaba open-source SparkDiffusion: 265x faster AI video on one RTX 5090 — lmoroney · 2026-10-01