Skip to content
localmodel.run

Text model · SmolLM2

SmolLM2 135M: RAM and VRAM requirements

SmolLM2 135M needs about 1 GB to run at Q4_K_M (the Q4_K_M GGUF file is ~0.105 GB to download; KV cache and overhead add the rest), or about 1 GB at Q8_0. The lightest hardware that runs it is Apple M1 (8GB).

SmolLM2 family · 0.135B params · released 2024-11-02 · 2.6M Ollama pulls.

Shopping for hardware? See what runs SmolLM2 135M →

License Apache-2.0 · Commercial OK ↓ 2M/mo ♥ 385 on HuggingFace
Q4_K_M GGUF
0.105 GB
Q8_0 GGUF
0.145 GB
Memory @ Q4 (4k)
~1 GB
Context
2 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?

SmolLM2 135M 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 0.9 GB usable 10.5 GB
See the full breakdown
$ollama run smollm2:135m

Memory breakdown

Weights (Q4_K_M)0.105 GB
+
KV cache (4k)0.1 GB
+
Overhead0.8 GB
=
Total1 GB

How context length changes it

4k context ~1 GB 32k context ~1.6 GB 128k context ~3.7 GB

SmolLM2 135M is small enough that the context window is the thing to size for: 128k needs ~3.7 GB versus ~1 GB at 4k, so pick the window you actually use.

Speed drops too: every token re-reads the KV cache, so at 128k context SmolLM2 135M generates at roughly ~4% of its short-context speed. How this is estimated.

Quantization sizes

GGUF quantson disk
Quantization Size on disk
Q2_K 0.1 GB est
Q3_K_M 0.1 GB est
Q4_K_M (default) 0.105 GB
Q5_K_M 0.1 GB est
Q6_K 0.1 GB est
Q8_0 0.145 GB
FP16 0.3 GB est

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

Run it

Ollama
$ ollama run smollm2:135m
llama.cpp
$ llama-cli -hf bartowski/SmolLM2-135M-Instruct-GGUF:Q4_K_M
LM Studio
$ lms get bartowski/SmolLM2-135M-Instruct-GGUF

Which devices can run SmolLM2 135M?

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 up Gemma 3 270M 270M

Similar models

Head-to-head

FAQ

How much VRAM or RAM does SmolLM2 135M need?

At Q4_K_M, SmolLM2 135M needs about 1 GB (weights ~0.105 GB + KV cache + overhead) at a 4k context. At Q8_0 budget ~1 GB.

What is the Q4_K_M GGUF file size of SmolLM2 135M?

The Q4_K_M GGUF file is about 0.105 GB to download, and the Q8_0 GGUF is about 0.145 GB. That is the weights file on disk; to run it you also need room for the KV cache and overhead, so budget ~1 GB of memory at Q4_K_M.

Can SmolLM2 135M run on a laptop?

Yes, SmolLM2 135M fits on a 16 GB machine at Q4_K_M and runs on Apple Silicon or a 12 GB+ GPU comfortably.

Can I use SmolLM2 135M commercially?

Yes. SmolLM2 135M is licensed Apache-2.0, which permits commercial use.

Understand the numbers

Short guides to the ideas behind SmolLM2 135M's memory and quant figures.

Hugging Face's smallest SLM. 135M params, fits in ~105MB at Q4_K_M. Designed for microcontrollers and on-device inference. Trained on 2T tokens (FineWeb-Edu, DCLM, The Stack).

Sources

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