Open-weights diffusion language model from DiffuSpace AI with 35B total and 3B 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.
Clover (DiffuSpace Clover 35B-A3B) is a diffusion-based language model from DiffuSpace AI. It has 35 billion total parameters with about 3 billion active per token, which indicates a mixture-of-experts design. It is published as a preview on the Hugging Face blog.
According to a YouTube review by Matthew Berman, the model scores well against similar-size models. In a computer-use demo it finished a task in 169 seconds, while Qwen 3.6 Flash reached its step limit at 578 seconds without completing it. These are third-party demo observations, not independent benchmarks.
The weights are described as open. The license, exact release date, context length and supported modalities were not available in the material provided, so those fields are left unknown or null. The "preview" label suggests the release may still change.
Watch the moment: the exact second top creators show it. Videos load only when you press play, via youtube-nocookie.com.
Open-weights diffusion language model with 35B total and 3B active parameters. It scores well against similar-size models. In a computer-use demo it finished a task in 169 seconds while Qwen 3.6 Flash hit its step limit at 578 seconds without completing.
Diffusion language model derived from Qwen 3.6 35B that generates blocks of tokens in parallel. Strong on agentic and tool-use tasks with favorable speed; full model 69GB.
Closest neighbours by meaning, not just by category.
Model · Google
Google's open-weights 26B diffusion language model that drafts blocks of text in parallel.
Model · Alibaba
Alibaba's Qwen3.8 Flash Next: a 125B-parameter MoE language model with about 6B active parameters, previewing Qwen 4 design.
Model · Amazon
Amazon's looped diffusion language models (1.7B and 8B) with adaptive per-token computation, built on Qwen3.
Model · Princeton
Fully open 3B text-to-image diffusion model from Princeton with released weights, training code, and data pipeline.
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
35B Mixture-of-Experts open-weights model from InternScience for long-horizon agentic search, engineering and tool use.
