Illustration, AI-generated from the story text. Tap a panel to read the story through it.

1 / 4DiffuSpace releases a 35-billion-parameter diffusion language model converted from Qwen×

The release

DiffuSpace posted Clover-35B-A3B-Preview on Hugging Face on October 24th under Apache 2.0, a mixture-of-experts model with about 3 billion active parameters converted from Qwen3.6-35B-A3B.

How it writes

Instead of writing one token at a time, it fills 32-token blocks of masked text in parallel.

Specs and gaps

It handles 128,000 tokens of context and needs about 69 GB of GPU memory. DiffuSpace calls it the strongest open diffusion language model, but its benchmarks appear only as an image and it gives no speed figures, which are the reason to use diffusion at all.

The cheap path

The cheap path it shows is converting an existing open model into a parallel decoder instead of training one from scratch.

Sources and the full story

‹›

DiffuSpace releases a 35-billion-parameter diffusion language model converted from Qwen

DiffuSpace posted Clover-35B-A3B-Preview on Hugging Face on October 24th under Apache 2.0, a mixture-of-experts model with about 3 billion active parameters converted from Qwen3.6-35B-A3B. Instead of writing one token at a time, it fills 32-token blocks of masked text in parallel. It handles 128,000 tokens of context and needs about 69 GB of GPU memory.

DiffuSpace calls it the strongest open diffusion language model, but its benchmarks appear only as an image and it gives no speed figures, which are the reason to use diffusion at all. The cheap path it shows is converting an existing open model into a parallel decoder instead of training one from scratch. SourcesA

Sources

• AClover-35B-A3B-Preview (DiffuSpace, Hugging Face)huggingface.co

This story ran in the Morning Brief of October 31st 2026.