What are the hardware requirements for LFM2.5 1.2B Thinking?

LFM2.5 1.2B Thinking requires approximately 2.3 GB of VRAM and supports a context window of 125K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed LFM2.5 1.2B Thinking and how large is it?

LFM2.5 1.2B Thinking is developed by LiquidAI, a model with 1.2b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

Where can I download or evaluate LFM2.5 1.2B Thinking?

LFM2.5 1.2B Thinking is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

LFM2.5 1.2B Thinking by LiquidAI — VRAM 2.3GB, 125K context

Name: LFM2.5 1.2B Thinking
Author: LiquidAI

LFM2.5 1.2B Thinking is an open-source language model by LiquidAI. Features: 1.2b LLM, VRAM: 2.3GB, Context: 125K, License: other, LLM Explorer Score: 0.32.

Arxiv:2511.23404 Ar Base model:finetune:liquidai/l... Base model:liquidai/lfm2.5-1.2... Conversational De Deploy:azure Edge En Endpoints compatible Es Fr Ja Ko Lfm2 Lfm2.5 Liquid Region:us Safetensors Zh

Model Card on HF 🤗: https://huggingface.co/LiquidAI/LFM2.5-1.2B-Thinking

LFM2.5 1.2B Thinking Parameters and Internals

LLM Name	LFM2.5 1.2B Thinking
Repository 🤗	https://huggingface.co/LiquidAI/LFM2.5-1.2B-Thinking
Base Model(s)	LiquidAI/LFM2.5-1.2B-Base LiquidAI/LFM2.5-1.2B-Base
Model Size	1.2b
Required VRAM	2.3 GB
Updated	2026-06-07
Maintainer	LiquidAI
Model Type	lfm2
Model Files	2.3 GB
Supported Languages	en ar zh fr de ja ko es
Model Architecture	Lfm2ForCausalLM
License	other
Context Length	128000
Model Max Length	128000
Transformers Version	4.57.2
Tokenizer Class	PreTrainedTokenizerFast
Padding Token	<\|pad\|>
Vocabulary Size	65536

Rank the LFM2.5 1.2B Thinking capabilities

Have you tried this model? Rate its performance — this feedback helps the ML community find the right model for their needs.

Instruction Following and Task Automation
Factuality and Completeness of Knowledge
Censorship and Alignment
Data Analysis and Insight Generation
Text Generation
Text Summarization and Feature Extraction
Code Generation
Multi-Language Support and Translation

Quantized Models of the LFM2.5 1.2B Thinking

Model	Likes	Downloads	VRAM
LFM2.5 1.2B Thinking 6bit	1	944	1 GB
LFM2.5 1.2B Thinking MLX 4bit	11	343	0 GB
LFM2.5 1.2B Thinking MLX 8bit	15	294	1 GB
LFM2.5 1.2B Thinking MLX 6bit	5	92	1 GB
LFM2.5 1.2B Thinking MLX 5bit	5	86	0 GB
LFM2.5 1.2B Thinking 8bit	2	59	1 GB

Best Alternatives to LFM2.5 1.2B Thinking

Best Alternatives	Context / RAM	Downloads	Likes
LFM2.5 1.2B Instruct	125K / 2.3 GB	575331	647
LFM2.5 1.2B JP 202606	125K / 2.3 GB	24006	69
LFM2.5 1.2B JP 202606 ONNX	125K / GB	9659	6
LFM2.5 1.2B Base	125K / 2.3 GB	22738	137
LFM2 1.2B	125K / 2.3 GB	147917	357
Catmind 1.2B	125K / 2.3 GB	433	4
LFM2.5 1.2B Instruct	125K / 2.3 GB	22673	8
...fm2.5 1.2B Instruct Manumit V1	125K / 2.3 GB	139	0
...fm2.5 1.2B Thinking Manumit V1	125K / 2.3 GB	138	0
...M2.5 1.2B Thinking Abliterated	125K / 2.4 GB	278	0

Note: green Score (e.g. "73.2") means that the model is better than LiquidAI/LFM2.5-1.2B-Thinking.