Skip to content
localmodel.run

Text model · LFM

LF LFM3-VL-2B: RAM and VRAM requirements

LFM3-VL-2B needs about 2.5 GB to run at Q4_K_M (the Q4_K_M GGUF file is ~1.38 GB to download; KV cache and overhead add the rest), or about 3.2 GB at Q8_0. The lightest hardware that runs it is Nvidia GeForce RTX 2060 (6GB).

LFM family · 1.92B params · released Oct 2026.

Shopping for hardware? See what runs LFM3-VL-2B →

Vision · accepts images License LFM Open License v1.0 · Conditional ↓ 9.1K/mo ♥ 195 on HuggingFace
Q4_K_M GGUF
1.38 GB
Q8_0 GGUF
2.07 GB
Memory @ Q4 (4k)
~2.5 GB
Context
125 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?

LFM3-VL-2B runs on 49 of 49 tracked devices at Q4_K_M.

49 run well 0 tight fit 0 too small
Fit check Q4_K_M
Yes, it runs fast
needs 2.5 GB usable 10.5 GB
See the full breakdown
$ollama run lfm3-vl:2b

Memory breakdown

Weights (Q4_K_M)1.38 GB
+
KV cache (4k)0.3 GB
+
Overhead0.8 GB
=
Total2.5 GB

How context length changes it

4k context ~2.5 GB 32k context ~4.9 GB 128k context ~12.8 GB

LFM3-VL-2B is small enough that the context window is the thing to size for: 128k needs ~12.8 GB versus ~2.5 GB at 4k, so pick the window you actually use.

Speed drops too: every token re-reads the KV cache, so at 128k context LFM3-VL-2B generates at roughly ~12% of its short-context speed. How this is estimated.

Quantization sizes

GGUF quantson disk
Quantization Size on disk
Q2_K 0.8 GB est
Q3_K_M 0.9 GB est
Q4_K_M (default) 1.38 GB
Q5_K_M 1.4 GB est
Q6_K 1.6 GB est
Q8_0 2.07 GB
FP16 3.58 GB

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

Run it

Ollama
$ ollama run lfm3-vl:2b
llama.cpp
$ llama-cli -hf LiquidAI/LFM3-VL-2B-GGUF:Q4_K_M
LM Studio
$ lms get LiquidAI/LFM3-VL-2B-GGUF

Which devices can run LFM3-VL-2B?

NVIDIA GPUs

Apple Silicon Macs

RAM-only laptops

iPhone & iPad

Android

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 LFM3-1.7B 1.73B Next size up Sarvam-1 2B 2B

Similar models

Head-to-head

FAQ

How much VRAM or RAM does LFM3-VL-2B need?

At Q4_K_M, LFM3-VL-2B needs about 2.5 GB (weights ~1.38 GB + KV cache + overhead) at a 4k context. At Q8_0 budget ~3.2 GB.

What is the Q4_K_M GGUF file size of LFM3-VL-2B?

The Q4_K_M GGUF file is about 1.38 GB to download, and the Q8_0 GGUF is about 2.07 GB. That is the weights file on disk; to run it you also need room for the KV cache and overhead, so budget ~2.5 GB of memory at Q4_K_M.

Can LFM3-VL-2B run on a laptop?

Yes, LFM3-VL-2B fits on a 16 GB machine at Q4_K_M and runs on Apple Silicon or a 12 GB+ GPU comfortably.

Can I use LFM3-VL-2B commercially?

Conditionally. LFM Open License v1.0: free for organizations below $10M annual revenue; larger organizations require separate terms.

Understand the numbers

Short guides to the ideas behind LFM3-VL-2B's memory and quant figures.

Liquid vision-language model. Measured official GGUF totals include the 378,996,032-byte BF16 vision projector: Q4_K_M 1,478,623,840 bytes, Q8_0 2,225,234,528 bytes, BF16 3,845,622,368 bytes, converted to GiB. The vendor text config specifies 128,000 tokens (125 Ki tokens), despite the prose 128K label. Text and image inputs; memory includes the complete encoder. Runtime memory and speed remain estimates.

Sources

Catalog updated 2026-11-01. Memory figures are estimates. See methodology.