3B-parameter open model from Nanbeige using a looped transformer, aimed at agentic and reasoning tasks.
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.
Nanbeige4.2-3B is a small language model from Nanbeige, published on ModelScope. According to a YouTube creator's summary, it uses a looped transformer architecture that reuses layers, and it is reported to outperform larger Gemma 4 and Qwen 3.5 models on agentic and reasoning benchmarks. The model file is reported to be about 8 GB.
The ModelScope page describes it as an open-source model. The scraped page text did not include a model card, so details such as context length, training data, license and benchmark numbers could not be confirmed here and should be checked on the model page.
With 3B parameters it is small enough to run locally on consumer hardware, which suits on-device or low-cost agentic and reasoning experiments. Benchmark claims come from a third-party video summary and have not been independently verified.
Watch the moment: the exact second top creators show it. Videos load only when you press play, via youtube-nocookie.com.
3B agentic model using a looped transformer that reuses layers, outperforming larger Gemma 4 and Qwen 3.5 models on agentic and reasoning benchmarks. Model is about 8 GB.
Closest neighbours by meaning, not just by category.
Model · Nanbeige
33B dense open-weights LLM from Nanbeige, tuned for agentic coding and professional cowork tasks.
Model · OpenBMB
Dense 2B-parameter open-weights language model from OpenBMB for on-device and edge deployment, Apache-2.0 licensed.
Model · Alibaba
Small 0.8B-parameter open-weights language model in Alibaba's Qwen 3.5 family, aimed at lightweight and on-device use.
Model · OpenBMB
1B-parameter dense open-weights language model from OpenBMB for on-device use, with hybrid thinking and 128K context.
Model · Alibaba
Alibaba's Qwen 3.5 large open-weights language model (~122B parameters) from the Qwen family.
Model · Alibaba
Alibaba's open-weights 2.4T-parameter MoE language model (95B active) from the Qwen3.8 family, with a hosted Max version.
