Home / Models / LFM2.5 8B A1B
Liquid

LFM2.5 8B A1B

LFM2.5 8B A1B is an open-weights 8B-parameter model by Liquid. It holds an Intelligence score of 17.9 on the local.ai leaderboard (2026-08 snapshot).

MoEedgeopen weights
17.9
local.ai Intelligence
Parameters
8B
MoE · 1B active parameters
Benchmarks (local.ai snapshot)
τ²-bench
94%
GAIA
9%
GDPval
10%
HuggingFace metadata · updated 2026-08-04
repo
LiquidAI/LFM2.5-8B-A1B
license
other
downloads
172,704
likes
723
Official HuggingFace repoOllamaCompare on the savings tool

Metadata refreshed from the HuggingFace API (snapshot 2026-08-15T15:53:42.405Z). Context length and long-form descriptions are the next pipeline stage.

Does it fit your hardware?

Detect or select your GPU first. Green = weights + context headroom, yellow = tight, grey = does not fit.

Q4_K_M 4.6GQ6_K 7.2GQ8_0 8.8GFP16 16G
Quantization≈ File sizeNote
IQ2_XXS2.2 GBExtreme compression, largest quality loss
IQ3_XXS2.9 GBVery small
Q3_K_M3.7 GBSmall
NVFP44 GBNVIDIA 4-bit
Q4_K_S4.3 GBCommon
Q4_K_M4.6 GBMost common sweet spot
Q5_K_M6.1 GBBalanced
Q6_K7.2 GBNear-lossless
Q8_08.8 GBNear-lossless
FP1616 GBOriginal weights

Actual files on HuggingFace

FileSize
model.safetensors15.77 GB
GGUF: LiquidAI/LFM2.5-8B-A1B-GGUF · 148,387 downloads
FileSize
LFM2.5-8B-A1B-BF16.gguf15.78 GBdownload
LFM2.5-8B-A1B-F16.gguf15.78 GBdownload
LFM2.5-8B-A1B-Q4_0.gguf4.51 GBdownload
LFM2.5-8B-A1B-Q4_K_M.gguf4.80 GBdownload
LFM2.5-8B-A1B-Q5_K_M.gguf5.62 GBdownload
LFM2.5-8B-A1B-Q6_K.gguf6.48 GBdownload
LFM2.5-8B-A1B-Q8_0.gguf8.39 GBdownload
GGUF: unsloth/LFM2.5-8B-A1B-GGUF · 23,588 downloads
FileSize
LFM2.5-8B-A1B-BF16.gguf15.78 GBdownload
LFM2.5-8B-A1B-MXFP4_MOE.gguf4.98 GBdownload
LFM2.5-8B-A1B-Q8_0.gguf8.39 GBdownload
LFM2.5-8B-A1B-UD-IQ1_M.gguf2.40 GBdownload
LFM2.5-8B-A1B-UD-IQ2_M.gguf2.57 GBdownload
LFM2.5-8B-A1B-UD-IQ2_XXS.gguf2.52 GBdownload
LFM2.5-8B-A1B-UD-IQ3_S.gguf3.33 GBdownload
LFM2.5-8B-A1B-UD-IQ3_XXS.gguf3.05 GBdownload
LFM2.5-8B-A1B-UD-IQ4_NL.gguf4.06 GBdownload
LFM2.5-8B-A1B-UD-IQ4_XS.gguf3.97 GBdownload
LFM2.5-8B-A1B-UD-Q2_K_XL.gguf2.72 GBdownload
LFM2.5-8B-A1B-UD-Q3_K_M.gguf3.67 GBdownload

What you save self-hosting this model

Uses the model's own API price when it exists; otherwise the closest commercial equivalent.

Waiting for hardware detection…

Related models

Liquid
Laguna S 2.1
Intelligence 71.6
Liquid
Laguna XS 2.1
Intelligence 58.8

RunLocal presents third-party leaderboard data (local.ai / Exo Labs, 2026-08) with its own fit and cost estimates. Verify pricing, licenses and file sizes at the source before deploying a model.