---
title: "A 367,000-Parameter Neural Network Living on a Floppy Disk: Tiny 2.1 Runs on Pentium and Speaks Aloud"
description: "Developer WollyMonjac has introduced Tiny 2.1, a 367k-parameter transformer running in DOS on a Pentium. The entire model takes up 1.44 MB on a floppy disk, requires 540 KB of RAM, speaks aloud, and is distributed under the MIT license."
date: 2026-11-22T14:14:41.877Z
lang: en
url: https://xab.info/en/posts/neural-network-tiny-2-1-floppy-dos-pentium
tags: [neural-network, dos, retro-computing, tiny-2-1, transformer, floppy-disk, open-source, mit-license]
author: "Alexander Buran"
publisher: "XAB.info"
---

# A 367,000-Parameter Neural Network Living on a Floppy Disk: Tiny 2.1 Runs on Pentium and Speaks Aloud

![Old PC with CRT monitor and a holographic brain hovering over a floppy disk on a wooden desk](https://xab.info/media/2026/11/22/neyroset-tiny-2-1-na-diskete-dos-pentium/neyroset-tiny-2-1-na-diskete-dos-pentium-1.webp)

## Key Points

- Tiny 2.1 is a ~367k-parameter transformer written in Turbo C 2.0, running in DOS on Intel 486DX/Pentium hardware
- The entire project fits on a 1.44 MB floppy disk and requires ~540 KB of RAM to run
- The model is trained on ~3,500 facts, supports voice-synthesized answers, and allows user fine-tuning
- The code is distributed under the MIT license, including commercial use

In a world where large language models demand server farms of thousands of graphic accelerators and terabytes of memory, a developer known as WollyMonjac has proven that artificial intelligence can fit onto a storage medium that still remembers the scent of magnetic tape. His project, Tiny 2.1, is a compact neural network designed to run in a DOS environment on computers equipped with Intel 486DX and Pentium processors. The entire software package, along with the model weights, fits onto a standard 1.44 MB 3.5-inch floppy disk, and requires approximately 540 KB of free RAM to run. For context, a typical early 1990s Pentium computer was equipped with 8–32 MB of RAM, making Tiny 2.1 fully compatible with the hardware of that era.

### A 367,000-Parameter Transformer in Turbo C 2.0

It is important to point out immediately: Tiny 2.1 is not a large language model in the conventional sense. It is a small transformer designed for question-answering, written in Turbo C 2.0—a compiler that was the industry standard in the DOS era. The model's architecture includes four layers, and all weights are stored in an 8-bit format, allowing them to fit into approximately 380 KB. The total number of parameters is about 367,000—orders of magnitude fewer than modern LLMs, but sufficient for specialized conversational interaction within a given knowledge base.

### Trained on 3,500 Facts with a 1994-Era Interface

The model has been trained on approximately 3,500 facts, with each fact included in the training set in multiple phrasing variations to improve question recognition robustness. The user interface completely replicates the aesthetics of old DOS applications: the screen displays a window for entering questions, dialogue history, and a neural network status monitoring panel. A unique highlight of the project is the ability to output answers via a synthesized robotic voice through the computer's built-in speaker—for an era when speech "from the speaker" felt like science fiction, this feature gives Tiny 2.1 the character of a true retro demo.

### Hallucinations, Fine-Tuning, and the MIT License

The developer honestly warns: beyond the knowledge embedded in the model, Tiny 2.1 is capable of confidently inventing answers, thus demonstrating the classic behavior of generative models encountering an unfamiliar query. However, the project includes a folder for additional training, allowing users to independently expand the fact base and adapt the neural network to specific topics. The program is distributed under the MIT license, which permits free use, modification, and distribution of the code, including commercial applications. This makes Tiny 2.1 not just a novelty, but a potential foundation for educational projects, retro computer clubs, and experiments with "lightweight" AI on constrained resources.

### Why This Matters in 2026

At a time when the AI industry is moving towards increasingly heavy and energy-intensive architectures, projects like Tiny 2.1 remind us of a fundamental principle: intelligence does not have to be infinitely large to be useful. For retro computing enthusiasts, students studying the basics of neural network architectures, and simply curious users who have a Pentium computer with a floppy drive sitting on a shelf, Tiny 2.1 provides a rare and accessible way to experience transformer models "in the flesh"—without cloud services, GPUs, or subscriptions. A floppy disk, 540 KB of memory, and a robotic voice from the speaker, and the neural network answers your questions.

## Fact-Check & Sources

- [Нейросеть на 367 тысяч параметров уместили на дискету — она отвечает голосом и требует всего 540 КБ оперативной памяти](https://www.ixbt.com/news/2026/11/22/neiroset-na-367-tysiac-parametrov-umestili-na-disketu-ona-otvecaet-golosom-i-trebuet-vsego-540-kb-operativnoi-pamiati.html) - The primary source contains all key facts: author WollyMonjac, model Tiny 2.1, 367k parameters, 4 layers, 8-bit weights ~380 KB, Turbo C 2.0, ~3,500 training facts, 540 KB RAM, 1.44 MB floppy disk, MIT license, voice synthesis, fine-tuning folder.

## FAQ

### What is Tiny 2.1 and what computers does it run on?

Tiny 2.1 is a compact question-answering transformer written in Turbo C 2.0. The model is designed to run in a DOS environment on computers with Intel 486DX and Pentium processors, requiring about 540 KB of free RAM.

### How many parameters does the model have and how much disk space does it take?

Tiny 2.1 has about 367,000 parameters, uses four layers, and stores weights in an 8-bit format (~380 KB). The entire software package along with files fits on a standard 1.44 MB floppy disk.

### What is the model trained on and can its knowledge base be expanded?

The model is trained on approximately 3,500 facts, each included in the dataset in multiple phrasing options. A folder for additional training has been added to the project, allowing users to independently expand the knowledge base.

### What license does Tiny 2.1 use?

The program is distributed under the MIT license, which permits using, modifying, and distributing the code, including for commercial purposes.