Skip to content
localmodel.run

Text model · LFM

LF LFM2.5 1.2B Thinking: RAM and VRAM requirements

LFM2.5 1.2B Thinking needs about 1.8 GB to run at Q4_K_M (the Q4_K_M GGUF file is ~0.68 GB to download; KV cache and overhead add the rest), or about 2.3 GB at Q8_0. The lightest hardware that runs it is Apple M1 (8GB).

LFM family · 1.17B params · released Jan 2026.

Shopping for hardware? See what runs LFM2.5 1.2B Thinking →

License LFM Open License v1.0 · Conditional ↓ 38.2K/mo ♥ 387 on HuggingFace
Q4_K_M GGUF
0.68 GB
Q8_0 GGUF
1.16 GB
Memory @ Q4 (4k)
~1.8 GB
Context
128 k

Q4_K_M and Q8_0 are GGUF download (file) sizes. Memory @ Q4 includes the KV cache and overhead, so it runs larger than the file on disk.

Will it run on your device?

LFM2.5 1.2B Thinking runs on 40 of 40 tracked devices at Q4_K_M.

40 run well 0 tight fit 0 too small
Fit check Q4_K_M
Yes, it runs fast
needs 1.8 GB usable 10.5 GB
See the full breakdown
$ollama run lfm2.5-thinking:1.2b

Memory breakdown

Weights (Q4_K_M)0.68 GB
+
KV cache (4k)0.3 GB
+
Overhead0.8 GB
=
Total1.8 GB

How context length changes it

4k context ~1.8 GB 32k context ~3.6 GB 128k context ~9.8 GB

LFM2.5 1.2B Thinking is small enough that the context window is the thing to size for: 128k needs ~9.8 GB versus ~1.8 GB at 4k, so pick the window you actually use.

Speed drops too: every token re-reads the KV cache, so at 128k context LFM2.5 1.2B Thinking generates at roughly ~8% of its short-context speed. How this is estimated.

Quantization sizes

GGUF quantson disk
Quantization Size on disk
Q2_K 0.5 GB est
Q3_K_M 0.6 GB est
Q4_K_M (default) 0.68 GB
Q5_K_M 0.8 GB est
Q6_K 1 GB est
Q8_0 1.16 GB
FP16 2.18 GB

Lower quant = smaller and faster, slightly lower quality. Q4_K_M is the common default.

Run it

Ollama
$ ollama run lfm2.5-thinking:1.2b
llama.cpp
$ llama-cli -hf LiquidAI/LFM2.5-1.2B-Thinking-GGUF:Q4_K_M
LM Studio
$ lms get LiquidAI/LFM2.5-1.2B-Thinking-GGUF

Which devices can run LFM2.5 1.2B Thinking?

Apple Silicon Macs

RAM-only laptops

iPhone & iPad

Android

NVIDIA GPUs

AMD GPUs

Same job, different size

Pick by what fits your memory: step down to free up VRAM, or step up if you have headroom.

Next size down Apple OpenELM 1.1B 1.1B Next size up MiniCPM-V 4.6 1.3B

Similar models

Head-to-head

FAQ

How much VRAM or RAM does LFM2.5 1.2B Thinking need?

At Q4_K_M, LFM2.5 1.2B Thinking needs about 1.8 GB (weights ~0.68 GB + KV cache + overhead) at a 4k context. At Q8_0 budget ~2.3 GB.

What is the Q4_K_M GGUF file size of LFM2.5 1.2B Thinking?

The Q4_K_M GGUF file is about 0.68 GB to download, and the Q8_0 GGUF is about 1.16 GB. That is the weights file on disk; to run it you also need room for the KV cache and overhead, so budget ~1.8 GB of memory at Q4_K_M.

Can LFM2.5 1.2B Thinking run on a laptop?

Yes, LFM2.5 1.2B Thinking fits on a 16 GB machine at Q4_K_M and runs on Apple Silicon or a 12 GB+ GPU comfortably.

Can I use LFM2.5 1.2B Thinking commercially?

Conditionally. LFM Open License: free under $10M revenue; larger companies need a commercial license.

Understand the numbers

Short guides to the ideas behind LFM2.5 1.2B Thinking's memory and quant figures.

The reasoning-tuned sibling of LFM2.5-1.2B: dense 1.17B, explicit chain-of-thought thinking mode, 128K context. Sizes measured from the vendor GGUF repo, cross-confirmed against the Ollama tag listing. LFM Open License v1.0.

Sources

Last validated 2026-08-03. Memory figures are estimates. See methodology.