Tencent releases WeMM-Embedding: 2B model beats 8B baselines
_reachsumit · x · 2026-08-26
Tencent released the WeMM-Embedding technical report, a family of universal multimodal embedding models supporting text, images, videos, visual documents, and arbitrarily interleaved inputs. The family includes 2B, 4B, and 9B variants. Training involves two stages: large-scale multimodal alignment followed by refinement using curated data, fine-grained relevance supervision, and cross-scale knowledge transfer. Evaluations show leading performance on public benchmarks: the 2B variant surpasses the previous 8B open-source baseline on MMEB-v2, while the 9B variant achieves a new state-of-the-art overall score of 80.6. The model has been deployed at scale in WeChat applications, including Channels, Official Accounts, Moments, and e-commerce services, yielding substantial gains.
Related event: Tencent Open-Sources WeMM-Embedding Multimodal Models, 9B Tops MMEB v2/v3(7 posts)→
More from Multimodal
- FLUX 3 Image Klein lands on fal: box-drawn layouts, multi-reference placement, style transfer — bfl_ai · 2026-10-16
- SGLang ships Day-0 support for FLUX 3 Image Klein: 10 reference images, BF16/FP8/NVFP4 variants — ying11231 · 2026-10-16
- Krea 3 Beta: 2K Edits in Under 8s, First Editing Model with LoRA Fine-tuning — testingcatalog · 2026-10-16
- Krea 3 Scores 4.81 on CPI-Bench, Second Only to GPT Image 2.5 — testingcatalog · 2026-10-16
- Krea 3 Pricing: $0.03 per 1K Image, 4x Faster Than Nano Banana Pro — testingcatalog · 2026-10-16
- Higgsfield teases Kling 4.0 rollout with a monkey-driven demo video — azed_ai · 2026-10-16