Model reference · open weights
LFM2.5 is an open-weight language model from LiquidAI. LFM2.5-8B-A1B (BF16) weighs 16.9 GB; the smallest configuration that runs it is 2× RTX 3060 12 GB.
Summary of the LiquidAI/LFM2.5-2.6B-GGUF model card, 2026-10-01. The estimate below is for another build of the family.
What it is
| Released by | LiquidAI |
|---|---|
| Released | 2026-08-01 |
| VRAM | 16.9 GB for the weights |
What it runs on
How much memory each request adds isn't estimated yet for this architecture. The weights need at least the cards below, plus room for the context.
| Card | Weights alone |
|---|---|
| RTX 3060 12 GB … RTX 4060 Ti 16 GB 2 smaller cards | does not fit |
| RTX 3090 24 GB | fits |
| RTX 4090 24 GB | fits |
| RTX 5090 32 GB | fits |
| L40S 48 GB | fits |
| A100 80 GB | fits |
| H100 80 GB | fits |
| RTX PRO 6000 Blackwell 96 GB | fits |
| DGX Spark (GB10) 128 GB unified | fits |
| H200 141 GB | fits |
| B200 180 GB | fits |
| 2× RTX 3060 12 GB tensor parallel | fits |
| 2× RTX 4060 Ti 16 GB tensor parallel | fits |
Builds
| Build | Params | Precision | Weights | Smallest setup |
|---|---|---|---|---|
| LFM2.5-2.6B-GGUF ↗ 7 builds: Q4_K_M 1.7 GB … F16 5.4 GB |
— | GGUF | 1.7 GB | RTX 3060 12 GB |
| LFM2.5-1.2B-Instruct ↗ | 1.2B | BF16 | 2.3 GB | RTX 3060 12 GB |
| LFM2.5-2.6B ↗ | 2.7B | BF16 | 5.4 GB | RTX 3060 12 GB |
| LFM2.5-350M ↗ | 354M | BF16 | 709 MB | RTX 3060 12 GB |
| LFM2.5-8B-A1B (above) ↗ | 8.5B | BF16 | 16.9 GB | 2× RTX 3060 12 GB |
| LFM2.5-230M ↗ | 230M | BF16 | 459 MB | RTX 3060 12 GB |
| LFM2.5-8B-A1B-MLX-4bit ↗ | 8.5B | MLX | 4.8 GB | RTX 3060 12 GB |
| LFM2.5-1.2B-Instruct-GGUF ↗ packaged by unsloth 19 builds: Q2_K 483 MB … BF16 2.3 GB |
— | GGUF | 731 MB | RTX 3060 12 GB |
From the model card
src="https://cdn-uploads.huggingface.co/production/uploads/61b8e2ba285851687028d395/2b08LKpev0DNEk6DlnWkY.png" alt="Liquid AI" style="width: 100%; max-width: 100%; height: auto; display: inline-block; margin-bottom: 0.5em; margin-top: 0.5em;" />
LFM2.5 is a new family of hybrid models designed for on-device deployment. It builds on the LFM2 architecture with extended pre-training and reinforcement learning.
Find more details in the original model card: https://huggingface.co/LiquidAI/LFM2.5-2.6B
Example usage with llama.cpp:
llama-cli -hf LiquidAI/LFM2.5-2.6B-GGUF --conversation \
--temp 0.1 --top-k 50 --repeat-penalty 1.1
The Quantization-Aware Distillation (QAD) checkpoint is available as
LFM2.5-2.6B-QAD-Q4_0.gguf.
This is distinct from the post-training-quantized LFM2.5-2.6B-Q4_0.gguf;
both use the GGUF Q4_0 format.
Quoted from the model card on Hugging Face. The full card is behind the Hugging Face link above.
How it works