Open MoE language model from Rednote (Xiaohongshu) with 208B total and 16B active parameters.
Updated EU status checked Last shown
Sources: the vendor's own pages (linked above) and creator videos at the linked timestamps. EU status is information, not legal advice; see how we check.
Dots3 Note (dots.3) is a large language model from Rednote (Xiaohongshu), published under the dots model line. According to a YouTube creator review, it is a mixture-of-experts model with 208B total parameters and 16B active per token, and the full checkpoint is about 577 GB (about 300 GB in FP8).
The same review says it performs close to much larger models and leads open models on ARC-AGI-2. These claims come from a third-party summary and were not confirmed on the official page text, which did not load in a readable form.
Details on release date, licence, pricing and hosted access were not available in the provided material. Weights are reported as open, but the repository and licence were not verified.
Watch the moment: the exact second top creators show it. Videos load only when you press play, via youtube-nocookie.com.
Open MoE model with 208B total and 16B active parameters; performs close to much larger models and leads open models on ARC-AGI-2. Full model 577 GB, FP8 300 GB.
Closest neighbours by meaning, not just by category.
Model · Alibaba
Alibaba's open-weights 2.4T-parameter MoE language model (95B active) from the Qwen3.8 family, with a hosted Max version.
Model · inclusionAI
Open-weights 560B hybrid-linear MoE reasoning model (25B active, 1M context) from inclusionAI, MIT licensed.
Model · Z.ai
Z.ai's open-weights 753B MoE language model, a post-trained GLM-5.2 aimed at coding, agentic tasks and cybersecurity.
Model · Alibaba
Alibaba's Qwen3.8 Flash Next: a 125B-parameter MoE language model with about 6B active parameters, previewing Qwen 4 design.
Model · AMD
AMD's fully open 16B-parameter Mixture-of-Experts language model with 2.8B active parameters, trained on Instinct GPUs.
Model · Baidu
Baidu's open-weights 280B-parameter sparse MoE base model (14B active) with 1M-token context, released pre-post-training.