Skip to content
NewA calmer, faster ZeroHour

Qwenputer compiles a RISC-V CPU into Qwen 3.5 4B weights

Original title: Qwenputer: A functioning, deterministic RISC-V CPU inside Qwen3.5

Hacker News · AIby csmantle ()9 min read1
AI summary · x-ai/grok-4.7Summaries by Vani AI

The experimental release treats Gated DeltaNet state as exact RAM and executes C and MicroPython during forward passes.

A researcher released Qwenputer, a stock Qwen 3.5 4B with extra layers that implement a 32-bit RISC-V core. The core uses Gated DeltaNet heads as addressable memory and runs as the model generates instruction-trace tokens. The author says it passes official riscv-tests, matches qemu, and ran 18 million instructions without error. Weights are on Hugging Face and ModelScope; code is on GitHub.

Unconfirmed: Correctness and the 18 million error-free instructions are the author's claims; the post is truncated.

Key facts
  • The CPU is RV32IM plus Zba and a small bignum unit.
  • It occupies 11 layers in front of Qwen's 32.
  • Qwen's original weights are unchanged; computer weights are added.
  • A driver named RvChat loads firmware into Gated DeltaNet state.
  • No training is used; a compiler writes the weights directly.
Explain this AI news with Vani AIWhat was announced, what is new and why it matters, in a new chat

From the original article

9 min read at divagr.com

A stock Qwen 3. 5 4B with a RISC-V CPU compiled into its weights, running on its own Gated DeltaNet layers as RAM. Qwenputer is a pretty fun experimental model.

Hacker News · AI

By csmantle · divagr.com

Read the full article at Hacker News · AI

A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.