Model

LFM2.5-8B-A1B-Base

lfm2-5-8b-a1b-base

JSON API Find compatible recipes

Record

Family: lfm
Architecture: dense
Parameters: 8
HF downloads / 30d: 1,735

Hugging Face model card

Identity

https://huggingface.co/LiquidAI/LFM2.5-8B-A1B-Base

Repository: LiquidAI/LFM2.5-8B-A1B-Base
Status: known
Link type: Exact Hub repository

Public Hugging Face repository confirmed by the Hub API.

Leaderboard scores

Scraped public quality scores joined to this model by its Hugging Face repository. They describe the listed variant, not any specific local quantization, and are never speed measurements.

Benchmark	Category	Score	Rank	Scored variant	Confidence
if-eval	instruction-following	91.84	13	LFM2.5-8B-A1B-GGUF	high
if-eval	instruction-following	79.44	160	LFM2.5-8B-A1B-GGUF	medium
ifbench	instruction-following	56.47	45	LFM2.5-8B-A1B-GGUF	high
ifbench	instruction-following	26	80	LFM2.5-8B-A1B-GGUF	medium
aime-2025	math	42.53	216	LFM2.5-8B-A1B-GGUF	high
aime-2025	math	20	241	LFM2.5-8B-A1B-GGUF	medium
aime-2026	math	50	112	LFM2.5-8B-A1B-GGUF	high
math-500	math	88.76	65	LFM2.5-8B-A1B-GGUF	high
math-500	math	74.8	101	LFM2.5-8B-A1B-GGUF	medium

Remaining fields

Identity, launch, related records, and measured speed are shown above. This is the rest of the normalized record.

architecture: dense
family: lfm
id: lfm2-5-8b-a1b-base
name: LFM2.5-8B-A1B-Base
params: 8
schema version: local-ai-registry/v1
url: huggingface.co/LiquidAI/LFM2.5-8B-A1B-Base ↗

downloads

all time: 12,910
captured at: 2026-08-31T15:38:41Z
last 30d: 1,735

Provenance & metadata (1)

source huggingface-api

Provenance & metadata (2)

facts

active params · provenance · captured at: 2026-08-27T06:04:13.773Z

active params · reason not-observed

active params · state unknown

provenance

captured at: 2026-08-27T06:04:13.773Z

sources

captured at	kind	url
2026-08-27T06:04:13.773Z	normalized-model	huggingface.co/LiquidAI/LFM2.5-8B-A1B-Base ↗