Look at the accuracy numbers and these things clearly don't know much yet, and I'm not about to hand one my hardest work. But you can see where it's going. As quantization and the MoE stuff keeps getting better, "good enough to just run on my own machine" keeps eating into more of what I'm currently paying a frontier lab for. Once a local model can handle like 80% of what I need, the math stops making sense for the subscription.
I just tested this on a bug fixing benchmark I'm working on.<p>It did not perform as well as I expected. Qwen2.5-Coder-3B (2 years old) outperformed it by a wide range -> fixing ~50% of bugs whereas this model only fixed ~12%.<p>Granted, it's not a coder specific model, but given its benchmark performance to Gemma models, and that it's two years newer, and that it's an MoE with 8B total params, I expected it to be more competitive.
Some of the coding-specific fine-tunes were really impressive boosts. Qwen2.5-3B-Instruct is also available [0] -- if it's not too much to ask, I'd be curious how more general models stack up in your benchmark?<p>[0] - <a href="https://huggingface.co/Qwen/Qwen2.5-3B-Instruct" rel="nofollow">https://huggingface.co/Qwen/Qwen2.5-3B-Instruct</a>
Anybody use their localcowork [1] before?
That is where the demo lives. Or not?<p>[1] <a href="https://github.com/Liquid4All/cookbook/tree/main/examples/localcowork" rel="nofollow">https://github.com/Liquid4All/cookbook/tree/main/examples/lo...</a>
This is super interesting, I'm particularly excited for this one as it may allow teams to scale this architecture for VLAs (vision language action models), and having sparser models means more real-time actions on a locally hosted model<p>demo link for anyone that wants to try this out <a href="https://playground.liquid.ai/chat?model=cmppnbgse000004l4bc8df3wx" rel="nofollow">https://playground.liquid.ai/chat?model=cmppnbgse000004l4bc8...</a>
Question: I have a dirty car and the car wash is just 50 meters away. Should I walk or drive to the carwash?<p>Answer: . . . . So, unless you have a compelling reason not to, walk to the car wash.
You did not express any desire to wash the car. After nudging, it's funnier:<p>> <i>Is my desire to make my car clean related to a car wash in any way? To driving to it?</i><p>> <i>Your desire to have a clean car isn’t tied to the existence or use of a car wash or to traveling to one. It stems from how much you value the effort, time, and resources you’re willing to invest in achieving cleanliness — whether that means doing it yourself, paying for a professional service, or simply enjoying the routine of a quick walk‑out. Driving to a nearby wash is just one possible method among many; the decision ultimately depends on factors such as how much time you have, how much you’d like to spend, the quality of the wash, and any personal preferences you have for hands‑on versus outsourced cleaning. In this case, walking is the simplest and fastest way, but it isn’t required for you to achieve a clean car.</i><p>Common sense is clearly there, but we should not underestimate the colossal heap of tacit <i>assumptions</i> that drive "obvious" decisions in our daily life.
I'm surprised these models haven't picked this up yet in the training data. Both Claude and ChatGPT missed that one when I posed the question to them last year.
ChatGPT still says walk but adds:<p>>The main reasons to drive such a short distance would be if you're bringing the car specifically to be washed, carrying something heavy, or the weather or walking conditions make it impractical.<p>>If your goal is to get your car washed, you'll need the car there—so driving makes sense. If you're just going to talk to someone at the car wash or check it out, walking is probably faster.
Why would a model know that one washes cars at a car wash? We don't clean our bodies at the body wash or clean the kitchen at the kitchen wash.
There's meaning in the term "car wash" that it understands. But I don't suspect anyone has taught it that for 99.9% of people, going to car wash ONLY means that you're going to wash your car and that it should make that implicit assumption.<p>What if you're the car wash owner? Or a maintenance technician? Pretty easy to just walk over there if you're just 50ft away.
to your point, when my Aussie friends first mentioned a "car park" to my north american born self, i wondered _momentarily_ what that was, then realized it's sort of a fun name for what i would call a parking lot.
Every model knows what a car wash is.
If it doesn't, what's the point using it? Trusting it with your workflows, your code?
I walk to the gas station more often than I drive there.
Yeah, but you are not washing yourself there, I suppose?<p>The whole twist here is that to wash your car, you need your car, so you cannot go by foot.
doesnt seem unreasonable.
Liquid does amazing work, but I kinda feel like they are overtraining their models. 38T tokens seems like a lot for an 8B model
Woah, chinchilla scaling is 20 x active_params. I think mistral was 2 x Chinchilla. This is 1800 x
The small models are getting really impressive.<p>I recently realized that Qwen3.5:4B is way more capable than I thought a model that size could be.<p>Combine that with the work Liquid puts into RL and fine tuning, and you get models that perform extremely well on minimal hardware.<p>Combine that with your own fine tuning, and you get a specialized tool that is fast, private, and doesn’t require internet connection.
Hmm, I asked it who made it, and it says Google?
Wow, this is fucking phenomenal. I fed it a long transcript asking it to create a summary and it executed it extremely well. For an 8B model this is quite impressive.
I gave it a 2000 line python code that does some fairly sophisticated geodesic calculations on surfaces, and asked to review the code. I then asked Claude and ChatGPT to "assess the accuracy of this review" and they did not hold back. That said, its a very small model, and very fast.
They seem… much better than all the models they compared against? What’s the catch?
Guess we can run this even on CPU!
Why does this not have (day-one) support for Ollama? The previous model is on there? Is it related to the ongoing refactor work or are people abandoning Ollama for other LLM engines?
Ollama is just llama.cpp but with their own interface ontop. Liquid does support llama.cpp, but Ollama is slow in updating its llama.cpp dependency.
It does, ollama pull maternion/lfm2.5
No vision support?
I really love how fast it is! Their press release comparing it on Strix Halo and M5 Max are impressive. It going twice as fast at GPU benchmarks even more so!
Homeopathic AI