# nesformer compiles an NES CPU into stock Qwen3.5 model weights

> Developer dbreunig releases nesformer, running a 6502 CPU inside Qwen3.5 model weights without training.

- Published: 2026-11-11T08:40:50Z
- Score: 6/10
- Tags: ai, emulator
- Port / platforms: NES -> PC (Windows), Linux / Steam Deck, Mac (Tool)
- Web page: https://emu-radar.customfw.xyz/news/nesformer-nes-cpu-compiled-into-qwen-weights/
- Sources: https://www.reddit.com/r/emulation/comments/1xjlnvz/nesformer_an_nes_cpu_compiled_into_the_weights_of/

Developer [dbreunig](https://github.com/dbreunig/nesformer) has released **nesformer**, a tool that compiles an NES's 6502 CPU, 2 KB of work RAM, and cartridge program ROM directly into the weights of a stock `Qwen3_5ForCausalLM` checkpoint. Building on [Divyansh Agrawal](https://divagr.com/blog/qwenputer)'s Qwenputer, the project computes every weight via a compiler rather than traditional training, using the model's Gated DeltaNet state as machine memory.

To try it, users need Python 3.11 or newer, NumPy, SciPy, Numba, and Pillow. Unzip the archives, check integrity with SHA256SUMS, and run NROM games like *Super Mario Bros.*

1. Set up a virtual environment and install dependencies via `pip install numpy scipy numba pillow`.
2. Run a test ROM with `python -m nesformer run roms/thwaite.nes --frames 600 --show`.

Execution takes one to six seconds per frame after an initial 20-second Numba compilation. Other cartridge boards stop with a message.
