SenseTime Open-Sources Visual Multimodal Model
HeyNayeem · x · 2026-07-14
SenseTime has open-sourced SenseNova-Vision-7B-MoT, featuring "one model covering major visual tasks." It can be controlled via language or visual prompts without needing task-specific heads. Officially, it outperforms Google DeepMind's Vision Banana on multiple vision benchmarks.
Comparisons provided include:
- RefCOCOg referring expression segmentation: 80.3 vs 73.8
- Cityscapes semantic segmentation: 71.2 vs 69.9
- NYUv2 depth estimation: 98.1 vs 94.8
- NYUv2 normal estimation: 14.4 vs 17.8
The forwarder emphasized that "new tasks are defined by instructions, not new heads," and included the code repository.
Related event: SenseTime open-sources SenseNova-Vision-7B-MoT unified vision model(6 posts)→
More from Models
- Google launches Gemini 3.5 Flash Cyber for CodeMender, with limited access for governments — GoogleAI · 2026-07-22
- Kimi K3 tops Gemini 3.6 Flash on four shared public benchmarks — ChrisGPT · 2026-07-22
- Google’s year-long pause in new base-model pretraining draws sharp criticism — teortaxesTex · 2026-07-22
- Current setup is 8,192 input tokens and 2,048 output tokens, with 8k/512 next — TheZachMueller · 2026-07-22
- Kimi K3 feels slower than K2.7, but stronger on long coding jobs and refactoring — Far-Presence2711 · 2026-07-22
- Poolside launches Laguna S 2.1 with 118B parameters and 8B active per token — Madisonkanna · 2026-07-22