Diffuspace/Clover-35B-A3B-Preview released
DiffuSpace has released its first model, Clover-35B-A3B-Preview, which combines a 35B-parameter mixture-of-experts backbone with approximately 3B active parameters per token. It features block diffusion generation and a 128K context window. This text-only model is derived from Qwen3.6-35B-A3B, maintaining its hybrid Gated DeltaNet (GDN) / full-attention MoE architecture with 40 layers, 256…
Key points
- DiffuSpace released Clover-35B-A3B-Preview
- Clover combines a 35B-parameter backbone with approximately 3B active parameters per token
- Context window is up to 128K tokens
Diffuspace/Clover-35B-A3B-Preview · Hugging Face
huggingface.co · 24 October 2026
Loading the full article…
This text was published by huggingface.co. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗
Coverage and discussion
1source- Reddit discussionreddit.com
The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.
More in Generative AI & Models
All →- NVIDIA releases open-weight Nemotron 3.5 Transcribe with multilingual transcription and speaker attribution · 7 src
- Ggml-org/llama.cpp merges PR #31283 adding video input for LFM3 models · 2 src
- Deepseek V4.1 Flash: 1,286 tokens/sec · 1 src
- inclusionAI releases Ling-3.1-flash-1019 with 1M-token window and 560B parameters · 3 src
- Researchers propose Scoped semantic invariance to improve LLM consistency · 2 src
Comments
via GitHub Discussions