LFM2-24B-A2B: Scaling Up the LFM2 Architecture • Hacker News Board

Re: LFM2-24B-A2B: Scaling Up the LFM2 Architecture

#11

Post by aziis98 » Sat, May 02, 2026, 1:54 PM UTC

I just tried the Q4_K_M variant of this [] and this is one of the first models that run at ~20tps on my laptop. I also tried it with some "hard" maths questions and it clearly knows much. Can't wait to try some local coding agent harnesses with it (I recently discovered kon [1] and dirac [2] and wanted to try them out)

The only thing I'm not sure about is if this model supports thinking or not.

[1]: https://github.com/0xku/kon

[2]: https://github.com/dirac-run/dirac

Re: LFM2-24B-A2B: Scaling Up the LFM2 Architecture

#12

Post by aziis98 » Sat, May 02, 2026, 2:08 PM UTC

meatmanek wrote: post #2

This model is pretty cool if you don't have a GPU - I was able to get I think 20 or 30 tokens per second on CPU (DDR4 ram) alone. (I don't remember if that was with q4 or q8.) Otherwise, if you have a GPU with more than like 4GB of VRAM, there are better models. Gemma4 and Qwen3.6 (or Qwen3.5 if you need the smaller dense models that haven't yet been released for 3.6) are a good place to start.

> I was able to get I think 20 or 30 tokens per second on CPU (DDR4 ram) alone

What are you using for inference? I have a recent intel laptop with 32GB of DDR5 and I am getting at most 25tps with the llama cpp vulkan backend (that is the fastest, I also tried sycl but it is a bit slower)

Re: LFM2-24B-A2B: Scaling Up the LFM2 Architecture

#13

Post by goldenarm » Sat, May 02, 2026, 2:20 PM UTC

Comparison with Qwen3.6 35B A3B:

- GPQA Diamond: 47.4% vs 84.1% for Qwen

- HLE: 4.4% vs 20.2% for Qwen

- AA Omniscience Accuracy: 6.4% vs 18.9% for Qwen

- AA Hallucination Rate: 30.0% vs 50.3% for Qwen

Re: LFM2-24B-A2B: Scaling Up the LFM2 Architecture

#14

Post by meatmanek » Sat, May 02, 2026, 6:33 PM UTC

aziis98 wrote: post #12

meatmanek wrote: post #2

This model is pretty cool if you don't have a GPU - I was able to get I think 20 or 30 tokens per second on CPU (DDR4 ram) alone. (I don't remember if that was with q4 or q8.) Otherwise, if you have a GPU with more than like 4GB of VRAM, there are better models. Gemma4 and Qwen3.6 (or Qwen3.5 if you need the smaller dense models that haven't yet been released for 3.6) are a good place to start.

> I was able to get I think 20 or 30 tokens per second on CPU (DDR4 ram) alone What are you using for inference? I have a recent intel laptop with 32GB of DDR5 and I am getting at most 25tps with the llama cpp vulkan backend (that is the fastest, I also tried sycl but it is a bit slower)

Ok, I double-checked, and I get 21-22tps with lmstudio-community/LFM2-24B-A2B-Q4_K_M.gguf running under LM Studio on my i5-12400 with 2x32GB sticks of DDR4 3200. This is with small context (just "Write me a poem about a language model named Liquid" in `lms chat`)

    Prediction Stats:
      Stop Reason: eosFound
      Tokens/Second: 21.10
      Time to First Token: 1.827s
      Prompt Tokens: 42
      Predicted Tokens: 187
      Total Tokens: 229

Re: LFM2-24B-A2B: Scaling Up the LFM2 Architecture

#15

Post by MarsIronPI » Sat, May 02, 2026, 7:46 PM UTC

trilogic wrote: post #6

Liquid AI have made some awesome models (especially the smaller ones, they are lightning fast). I wish they made a fast small size coder. Did a finetune distill of 0.8B myself and it is in fact working properly, coding like a 30B model, so I know it is possible. Anyway here you have the 24B parameters with 2B active: https://hugston.com/models/lfm2-24b-a2b-q4-k-m

That sounds pretty interesting. Did you publish a write-up anywhere? If not, could you say more about how you did the finetune? Which model did you fine-tune/distill, what datasets did you use?