---
title: "We’ve Found the Only Good AI Chatbot: One So Tiny It Fits on a Single MS-DOS Floppy Disk"
description: "A tiny AI model is so small that it can fit on a floppy disk and run on a computer old enough to still use one."
date: "2026-11-21"
modified: "2026-11-21"
authors:
  - name: "Frank Landymore"
    job_title: "Contributing Writer"
    link: "https://futurism.com/authors/flandymore.md"
url: "https://futurism.com/artificial-intelligence/ai-model-floppy-disk"
categories:
  - "Artificial Intelligence"
---

# We’ve Found the Only Good AI Chatbot: One So Tiny It Fits on a Single MS-DOS Floppy Disk

![A DOS floppy disk.](<https://futurism.com/wp-content/uploads/2026/11/good-llm-floppy-disk.jpg>)
*Shutterstock / Futurism*

At what size does a large language model stop being "large"?

The eagle eyes at *Hackaday* [spotted](<https://hackaday.com/2026/11/18/not-so-large-language-model-fits-on-a-floppy-disk/>) a diminutive AI model that, at just 380 KB in size, could fit on a single floppy disk and run on a computer old enough to still use one. In an age when trillion parameter AI models are basically [undoing decades worth of greenhouse gas reductions](<https://futurism.com/artificial-intelligence/data-centers-us-greenhouse-gas-reductions-emissions.md>) with their enormous power bills and [driving people into bouts of insanity](<https://futurism.com/chatgpt-mental-health-crises.md>) with their bewitchingly human-like responses, you could say this quaint little thing, by comparison, is the only good "LLM."

Don't expect Claude-like performance. The simple transformer model was trained on just 3,500 facts, and its capabilities are thin. It makes up answers pretty much all the time, since it doesn't actually "know" all that much about the world.

According to the [project page on GitHub](<https://github.com/WollyMonjac/Tiny-2.1>), it was written in Turbo C 2.0 and runs on MS-DOS, the ancient Microsoft operating system that was the go-to for personal computers throughout the 1980s and early 90s. The creator said they used Anthropic Fable 5.2 to write the model and Opus 5.5 to train it — so you are essentially getting a tiny AI model created by bigger AI models.

In all, it contains just 367,000 parameter, which is peanuts to the low-billions that your typical "lightweight" model has, and comprises 380 KB worth of 8-bit weights, the actual meat of the AI architecture. It also needs just 540 KB of memory to run, a necessity on an old operating system like DOS.

The interface is appropriately retro. A screenshot shows a typical example of the quality of the AI's responses: it correctly responds that whales are mammals, not fish, and that Alexander Graham Bell invented the telephone. But it also replies that 17 times 23 is 96, and gave an even more bizarre answer to "who is the king of france."

"It was Antarce Newton," the AI replied with a confidence level of 100 percent. We're not sure if there was ever a single person in human history named "Antarce Newton," but if this fun-sized model is saying that's the case, then let's just pretend its true.

**More on AI:** [*AI Is Breaking the Entire Structure of Software That Powers the World*](<https://futurism.com/artificial-intelligence/ai-breaking-software-powers-world.md>)

## Author
At Futurism, my work has often centered on bringing a sense of clarity and insight to complex topics ranging from the regulation of emerging technologies to the esoteric ideologies of Silicon Valley executives, while striving not to lose the poetic sense of awe inspired by often-obscure fields like astrophysics and quantum computing. I broke the story of CNET using AI to produce articles that turned out to be riddled with factual errors and plagiarism — a dam-breaking inflection point, as I've reported, that's inspired copycats and endless discourse while beguiling stakeholders ranging from tech giants to purveyors of spam around the web. My work at Futurism has been cited by publications including CBS News, the Los Angeles Times, Vice, Gizmodo, Engadget, the Verge, and Vanity Fair. I grew up in locales ranging from India to China, and now live in the exotic suburbs of Virginia. In my free time, I'm an avid reader of weird sci-fi literature, an aficionado of East Asian cinema, and, regrettably, a relapsed gamer. Allegedly, I’m working on a debut novel, currently untitled.

### Author social links  
[Bluesky](<https://bsky.app/profile/f-w-l.bsky.social>)

## More

- [Police Arrest Man With Almost Zero Resemblance to Actual Perpetrator Because AI Told Them To](<https://futurism.com/police-facial-recognition-nypd.md>)
- [OpenAI Warns Against Investing in Its Stock](<https://futurism.com/openai-warns-against-investing.md>)
- [Users Were So Addicted to GPT-4o That They Immediately Cajoled OpenAI Into Bringing It Back After It Got Killed](<https://futurism.com/users-addicted-gpt-4o-convinced-openai-bring-back.md>)
- [GPT-5 Is Turning Into a Disaster](<https://futurism.com/gpt-5-disaster.md>)
- [GPT-5 Launch Demo Plagued With Catastrophically Dumb Errors](<https://futurism.com/gpt-5-demo-dumb-errors.md>)
- [Man Follows ChatGPT’s Advice and Poisons Himself](<https://futurism.com/man-poisons-himself-chatgpt.md>)
- [Top AI Experts Concerned That OpenAI Has Betrayed Humankind](<https://futurism.com/openai-open-letter-experts-humankind.md>)
- [OpenAI Is Quietly Trying to Get More Money as It Burns Through Cash at a Staggering Pace](<https://futurism.com/openai-money-softbank-investors.md>)
