Very small: can run on the edge to allow something like a Raspberry Pi to make basic decisions for your appliance even if disconnected from the internet. Example: those are some time series parameters and instructions, decide if watering the plants or not; vision models that can watch a camera and transcribe what it is seeing in a basic way, ...

Small: runs in an average laptop not optimized for inference of LLMs, like Gemma 3 4B.

Medium: runs in a very high spec computer that people can buy for less than 5k. 30B, 70B dense models or larger MoEs.

Large: Models that big LLM providers sell as "mini", "flash", ...

Extra Large / SOTA: Gemini 2.5 PRO, Claude 4 Opus, ChatGPT O3, ...

• mnahkies

  today at 3:48 PM

  I'm not sure if you're implying that very small language models would be run in your raspberry pi example, but for use cases like the time series one, wouldn't something like an LSTM or TiDE architecture make more sense than a language model?

  These are typically small and performant both in compute and accuracy/utility from what I've seen.

  I think with all the hype at the moment sometimes AI/ML has become too synonymous with LLM

  • antirez

    today at 4:21 PM

    Sure if you have a specific need you can specialize some NN with the right architecture, collecting the data, doing the training several times, testing the performances, ... Or: you can download an already built LLM and write a prompt.

    • mnahkies

      today at 5:51 PM

      So one of the use cases we're serving in production is predicting energy consumption for a home. Whilst I've not tried, I'm very confident that providing an LLM the historical consumption and asking it to predict future consumption will under perform compared to our forecasting model. The compute required is also several orders of magnitude lower compared to an LLM

    • galangalalgol

      today at 5:05 PM

      What zero shot would you suggest for that task on an rpi? A temporal fusion thing?

      • antirez

        today at 5:29 PM

        The small gemma 3 and Qwen 3 models can do wonders for simple tasks as bag of algorithms.

        • galangalalgol

          today at 5:41 PM

          Those would use more ram than most rpi have wouldn't they? Gemma uses 4GB right?

          • antirez

            today at 6:47 PM

            Nope, gemma3 and qwen3 exist of many sizes, including very small ones, that 4-bit quantized can run on very small systems. Qwen3-0.6B, 1.7B, ... imagine if you quantize those to 4 bit. But there is the space for the KV cache, if we don't want to limit the runs to very small prompts.

          • nolist_policy

            today at 6:48 PM

            Gemma 3 4B QAT int4 quantized from bartowsky should barely fit in a 4GB Raspberry Pi, but without the vision encoder.

            However the brand-new Gemma 3n E2B and E4B models might fit with vision.

            • antirez

              today at 7:10 PM

              Yep, the Gemma 3 1B would be 815MB, with enough margin for a longer prompt. Probably more realistic.

  • greenavocado

    today at 4:11 PM

    He's talking about general purpose zero shot models.

• mnky9800n

  today at 4:42 PM

  Why in the world do you need such sophistication to know whether to water the plants or not?

  • collingreen

    today at 4:52 PM

    When you have a golden hammer everything starts to look like a nail

    • dainiusse

      today at 5:01 PM

      this

  • kovezd

    today at 4:52 PM

    There are places where: a) weather predictions are unreliable, b) there is scarcity of water. Just making the right decision on at what hour to water is a huge monthly saving of water.

    • 1over137

      today at 5:17 PM

      None of which need AI hype crap. Some humidity sensors, photosensors, etc. will do the job.

      • mnky9800n

        today at 8:24 PM

        I think there’s two schools of thought. The models will get so big everyone everywhere will use them for everything and they will make lots of money on api calls. The models will get cheaper and cheaper computationally on inference that implementing them on the edge will cost nothing and so an LLM will be in everything. Then every computational device will have one as long as you pay a license fee to the people who trained them.

      • kovezd

        today at 6:51 PM

        Need is a very strong word. We don't need a lot of we have today.

        But as a hobbyist I would prefer to program in an LLM than learn a bunch of algorithms, and sensor readings. It's also very similar to how I would think about it, making it easier to debug.

      • thenthenthen

        today at 7:11 PM

        Or a farmer

    • ithkuil

      today at 5:33 PM

      Does it have to be computed at the edge by every person?

      • kovezd

        today at 6:53 PM

        Just as the other comment "have to" is a very strong word. But there are benefits to it: a) adaptability to local weather patterns, b) no access to WiFi in large properties.

  • amelius

    today at 6:41 PM

    In this case, "sophistication" meaning throwing insane amounts of compute power and data at the problem? In older times we'd probably call that "brute forcing".

• SkiFire13

  today at 4:52 PM

  > Example: those are some time series parameters and instructions, decide if watering the plants or not

  How is that a "language model"?

  • tayo42

    today at 5:00 PM

    Is language model used to mean neural net, with transformers, attention that takes in a series of tokens and out outs a prediction as a value?

    Working with time series data would work in that case.

    • tough

      today at 5:16 PM

      https://github.com/google-research/timesfm

• layer8

  today at 4:27 PM

  For “very small”, I would add “can be passively cooled” as a criterion.

• oezi

  today at 7:10 PM

  How do we call the models beyond extra large which are so big they can't be served publicly because their inference cost is too high? Do such exist?

• lloydatkinson

  today at 7:41 PM

  > Example: those are some time series parameters and instructions, decide if watering the plants or not; vision models that can watch a camera and transcribe what it is seeing in a basic way, ...

  This is the problem I have with the general discourse of "AI" even on Hacker News, of all places. Everything you listed is not an example of a *language model*.

  All of those can either be implemented as a simple "if", decision tree, decision table, and finally actual ML in the example of cameras and time series predication.

  Using an LLM is not just ridiculous here but totally the wrong fit and a waste of resources.

I think of “fits on the overpowered M1/2/3/4 64GB MacBook Pro my employer gave me” as the dividing line. We’re getting to within spitting distance of models that can code well at that size.

• Maxious

  today at 3:47 PM

  https://mistral.ai/news/devstral and https://huggingface.co/nvidia/AceReason-Nemotron-14B were released in just the last couple of days and work in 24GB 4090 GPUs/32GB Macbook Pros just fine

• api

  today at 2:35 PM

  I want my next laptop to be the 128gb M series monster. That will run not quite frontier models but ones that are close in performance, and run them fast.

  • danielbln

    today at 4:30 PM

    And, also quite important, leave your system enough RAM to do anything else.

  • today at 2:43 PM

There is a "small language model", and then there is a "small LARGE language model". In late 2018, BERT (110 million params) would've been considered a "large" language model. A "small" LM would be some markov chain or a topic model (e.g. latent dirichlet allocation) - technically they would be considered generative language models since they learn joint distributions of params and data (words), and can then sample from that distribution. But today, we usually map "small" LMs to "small" LLMs, so in that sense a small LLM would be anything from BERT to around 3-4B params.

> Small models used to mean tiny. Now they mean "runs without drama."

Does this mean without a dedicated electric power plant?

I wanted to say "Right, big-sized. Do you want fries with that?", but I couldn't figure out how to work that in, so I won't say it.

Maybe we should appropriate the old DOS/x86 memory model names and give them “class-relative” sizes.

“tiny” can run on a microcontroller, “compact” on a Rpi, “small” on a phone, “medium” on a single GPU machine, “large” on AI class workstation hardware, and “huge” on a data center cluster.

On this topic, I've been wondering if models are capable of recommending other models for a given machine spec, for example: which model, if any, would be recommended for a laptop with a Ryzen 9 6000S and RTX 3060m (random spec).

I want to see more models that can be streamed to a browser and run locally via wasm. That would be my hope for small models. In the <100mb range.

• firejake308

  today at 4:05 PM

  After experimenting with 1B models, I am starting to think that any model with 1B parameters or less will probably lack a lot of the general intelligence that we observe in the frontier models, because it seems physically impossible to encode that much information into so few parameters. I believe that in the range of very small models, the winner will be models that are fine tuned to a small range of tasks or domains, such as a model that can translate between English and any other language, or a legal summarization model, etc.

• vindex10

  today at 4:59 PM

  Have you heard of Transformers.js? They are running onnx inside browser:

  https://huggingface.co/docs/transformers.js/en/index

• today at 2:41 PM

• relaxing

  today at 4:10 PM

  Why? Just so user data stays local?

  • dainiusse

    today at 5:02 PM

    Yes. And also, cost to run it.

A traditional Markov model trained (rather, just "fitted") on tokens or words is a small language model.

• GolDDranks

  today at 6:26 PM

  (To share a recent personal experience about Markov models: I bootstrapped recently a HMM with hand-assigned weights. It was around 15x15 class transitions, 225 weights. That's small. Or rather, microscopic. Then I ran it against real data, and picked up examples of wrong classifications, and made them auxillary training data. Of course, it was not a language model, language model is impossible to fit in such a small space. It was a model of transitions of chapter "types" in novels, where types are something like "Epilogue" , "Prologue", "Chapter 23", "Table of Contents", "Afterword" etc.)

The term is too overloaded.

I'll add one more: a LLM small enough that it can be trained from scratch on one A100 in 24 hours. Is it really small if it takes $10,000 to train? Or leave that term for $200 models?

Back to your definitions, there are sub-1B models people are using. I think I saw one in the 400-600M range for audio. Another person posted here a 100M-200M model for extracting data from web pages. We told them to just use a rules-based approach where possible but they believed the SLM worked better.

Then, there's projects like BabyLM that can be useful at 10M:

https://babylm.github.io/

• GardenLetter27

  today at 2:38 PM

  But you only have to train the foundational model once - so with open weights it's not really a problem.

  Maybe resources needed for fine-tuning would be nice to see.

  • nickpsecurity

    today at 8:13 PM

    Most have been trained on illegally-distributed, copyrighted works. They might output them, too. People might want untainted models. Additionally, some have weaknesses due to tokenizers, pre-training data, or moral alignment (political bias).

    For those reasons, users might want to train a new model from scratch.

    Researchers of training methods have a different problem. They need to see whether a new technique, like an optimization algorithm, gets better results. They try them more quickly with less money if they have small, training runs representative of what larger models do. If BabyLM-10M was representative, they could test each technique at the FLOPS/$ of a 10M model instead of a 1B model.

    So, both researchers and users might want new models trained from scratch. The cheaper to train, the better.

• today at 2:49 PM

I appreciate how it redefines “small” not by parameter count but by practical impact and deployability.

• lblume

  today at 3:35 PM

  I do not — parameter count is objective, practical impact depends on such a multitude of factors that any comparison becomes virtually meaningless.

  • kergonath

    today at 7:52 PM

    The standard for parameters count is rapidly evolving. Something large now will be small tomorrow, there is no point in using such a moving target as a criterion.

    • lblume

      today at 8:00 PM

      Sure, but nonetheless whether the model is called "small" at some time t should depend on the parameter count and t, not some arbitrarily specified metric of deployability.

These terms are all relative, but there's also "BabyLlama", which measures its parameter count in millions rather than billions.

This post is 100% rewritten or fully generated by gpt-4o. It has the gpt smell all over it.

• gwern

  today at 4:55 PM

  > In a world chasing ever-bigger models, small ones are quietly doing more with less—and that's exactly what makes them powerful.

  100%. It has enough technical details that maybe a human did something. But who knows.

• maksimur

  today at 5:54 PM

  Is there a problem with that? If so, what is it? I don't mind as long as it's not the boilerplate AI spits out by default.

  • stephantul

    today at 7:54 PM

    Nah not really, the information content is what counts of course. It’s just a bit cringe to see it happen.

This is also where MoE shines with a mixture of small and large language models.

How can a Large Language Model be a small language model?

• kelseyfrog

  today at 3:31 PM

  Because words are arbitrary. See Saussure.

• baq

  today at 4:01 PM

  Why wouldn’t there be any? Right now there are large large language models, medium large language models and small large language models. You can say there are also tiny large language models and extra large large language models. Nothing confusing about it.

• tialaramex

  today at 3:50 PM

  See also the Little Giant Girl who is part of The Sultan's Elephant and several other Royal de Luxe performances. She's clearly a little girl, but, she's also clearly a giant.

whatever fits into gaming GPU such as GeForce 3080

[dead]

Just ask my ex-wife!