Skip to content
Local AI Wizard
Your machine
Theme

SmolLM3 3B

Hugging Face · HuggingFaceTB/SmolLM3-3B

Q4 file 1.9 GB (Q4_K_M) · runs well from 6 GB · Weights license: Apache 2.0

Reported by the publisher33 resultsHugging Face's model card · read 2026-10-25 · not ranked · see them
Memory needed~3 GBQ4 weights + working room, before the machine's reserve
Parameters3 Bmanufacturer specification
Context65,536 tokensPublished for this model

Quality

No result in LiveBench 2026-06-25, LMArena 2026-10-22 Not in LiveBench 2026-06-25 or LMArena 2026-10-22

Reported by the publisher

33 results · Hugging Face's model card · read 2026-10-25
AIME 2025
9.3SmoLLM3-3B · No Extended Thinking · High school math competition · non-thinking mode
GSM-Plus
72.8SmoLLM3-3B · No Extended Thinking · Math problem-solving · non-thinking mode
LiveCodeBench v4
15.2SmoLLM3-3B · No Extended Thinking · Competitive programming · non-thinking mode
GPQA Diamond
35.7SmoLLM3-3B · No Extended Thinking · Graduate-level reasoning · non-thinking mode
IFEval
76.7SmoLLM3-3B · No Extended Thinking · Instruction following · non-thinking mode
MixEval Hard
26.9SmoLLM3-3B · No Extended Thinking · Alignment · non-thinking mode
BFCL
92.3SmoLLM3-3B · No Extended Thinking · Tool Calling · non-thinking mode
Global MMLU
53.5SmoLLM3-3B · No Extended Thinking · Multilingual Q&A · non-thinking mode
Every reported result33
AIME 2025
36.7SmoLLM3-3B · Extended Thinking · High school math competition · thinking mode
GSM-Plus
83.4SmoLLM3-3B · Extended Thinking · Math problem-solving · thinking mode
LiveCodeBench v4
30.0SmoLLM3-3B · Extended Thinking · Competitive programming · thinking mode
GPQA Diamond
41.7SmoLLM3-3B · Extended Thinking · Graduate-level reasoning · thinking mode
IFEval
71.2SmoLLM3-3B · Extended Thinking · Instruction following · thinking mode
MixEval Hard
30.8SmoLLM3-3B · Extended Thinking · Alignment · thinking mode
BFCL
88.8SmoLLM3-3B · Extended Thinking · Tool Calling · thinking mode
Global MMLU
64.1SmoLLM3-3B · Extended Thinking · Multilingual Q&A · thinking mode
HellaSwag
76.15SmolLM3-3B · English benchmarks · Reasoning & Commonsense · 0-shot
ARC-CF (Average)
65.61SmolLM3-3B · English benchmarks · Reasoning & Commonsense · 0-shot
Winogrande
58.88SmolLM3-3B · English benchmarks · Reasoning & Commonsense · 0-shot
CommonsenseQA
55.28SmolLM3-3B · English benchmarks · Reasoning & Commonsense · 0-shot
MMLU-CF (Average)
44.13SmolLM3-3B · English benchmarks · Knowledge & Understanding · 0-shot
MMLU Pro CF
19.61SmolLM3-3B · English benchmarks · Knowledge & Understanding · 0-shot
MMLU Pro MCF
32.70SmolLM3-3B · English benchmarks · Knowledge & Understanding · 0-shot
PIQA
78.89SmolLM3-3B · English benchmarks · Knowledge & Understanding · 0-shot
OpenBookQA
40.60SmolLM3-3B · English benchmarks · Knowledge & Understanding · 0-shot
BoolQ
78.99SmolLM3-3B · English benchmarks · Knowledge & Understanding · 0-shot
HumanEval+
30.48SmolLM3-3B · English benchmarks · Coding & math · 0-shot
MBPP+
52.91SmolLM3-3B · English benchmarks · Coding & math · 0-shot
MATH (4-shot)
46.10SmolLM3-3B · English benchmarks · Coding & math · 4-shot
GSM8k (5-shot)
67.63SmolLM3-3B · English benchmarks · Coding & math · 5-shot
Ruler 32k
76.35SmolLM3-3B · English benchmarks · Long context · 0-shot
Ruler 64k
67.85SmolLM3-3B · English benchmarks · Long context · 0-shot
Ruler 128k
61.03SmolLM3-3B · English benchmarks · Long context · 0-shot

Values as printed on the card, with the publisher's own settings (card section: Evaluation › Instruction Model › No Extended Thinking; Evaluation › Instruction Model › Extended Thinking). Not comparable across publishers and not used in any ranking. Model card

Run it

Run it

ollama run hf.co/bartowski/HuggingFaceTB_SmolLM3-3B-GGUF:Q4_K_M
  1. Install Ollama
  2. Run the command above
  3. Ask: Explain what a context window is, in two sentences.

Runs on

Fast memory holds it 118

The planning estimate fits the memory the model can use quickly: graphics card VRAM, or unified memory within the macOS GPU share.

Show all 118

Within memory, outside the fast part 1

Within the machine's memory, outside its fast part: on a graphics card the layers that do not fit run on the CPU from PC RAM (the listed RAM, or 32 GB assumed); on a Mac, above the macOS default GPU share. Not a speed figure.

SmolLM3 3B is a text model from Hugging Face with 3B parameters. It was released on July 8, 2025 and reads up to 65,536 tokens of context. Its Q4 weights plus working room come to about 3 GB before the machine's reserve; the smallest reference machines whose fast memory holds that are Mac mini M2 · 8 GB unified (~5 GB with its reserve) and RTX 4060 · 8 GB VRAM (~5 GB with its reserve). We collected 27 downloadable GGUF and MLX files from 3 publisher sources; the most downloaded GGUF source is bartowski's. It has no result in LiveBench 2026-06-25, so no quality score is shown.

Hardware profiles grouped by the Q4 planning estimate on each machine (dedicated GPUs reserve 2 GB; unified and system memory 25%, 2 to 8 GB): the fast memory holds it, it is within memory but outside the fast part (layers offloaded to PC RAM, or above the macOS GPU share), or the machine is one memory tier short. Each card opens its hardware profile. Not a measured fit or speed.

On the reference machines

SmolLM3 3B on the reference machines
Where Q4 sits
Mac mini M2 · 8 GB unifiedFast memory holds itQ8_0GGUF · 3.3 GB · needs ~7 GB16,384 tokens~5 GBFast memory holds itQ8_0GGUF · 3.3 GB · needs ~7 GB16,384 tokens
Mac mini M4 · 16 GB unifiedFast memory holds itQ8_0GGUF · 3.3 GB · needs ~9 GB65,536 tokens~7 GBFast memory holds itQ8_0GGUF · 3.3 GB · needs ~9 GB65,536 tokens
Mac mini M4 · 24 GB unifiedFast memory holds itQ8_0GGUF · 3.3 GB · needs ~11 GB65,536 tokens~9 GBFast memory holds itQ8_0GGUF · 3.3 GB · needs ~11 GB65,536 tokens
Mac mini M4 · 32 GB unifiedFast memory holds itQ8_0GGUF · 3.3 GB · needs ~13 GB65,536 tokens~11 GBFast memory holds itQ8_0GGUF · 3.3 GB · needs ~13 GB65,536 tokens
Mac mini M5 Pro · 64 GB unifiedFast memory holds itQ8_0GGUF · 3.3 GB · needs ~13 GB65,536 tokens~11 GBFast memory holds itQ8_0GGUF · 3.3 GB · needs ~13 GB65,536 tokens
Mac Studio M5 Max · 128 GB unifiedFast memory holds itQ8_0GGUF · 3.3 GB · needs ~13 GB65,536 tokens~11 GBFast memory holds itQ8_0GGUF · 3.3 GB · needs ~13 GB65,536 tokens
Mac Studio · M5 Ultra · 36-core CPU / 80-core GPU · 512 GB unifiedFast memory holds itQ8_0GGUF · 3.3 GB · needs ~13 GB65,536 tokens~11 GBFast memory holds itQ8_0GGUF · 3.3 GB · needs ~13 GB65,536 tokens
RTX 4060 · 8 GB VRAMFast memory holds itQ8_0GGUF · 3.3 GB · needs ~7 GB16,384 tokens~5 GBFast memory holds itQ8_0GGUF · 3.3 GB · needs ~7 GB16,384 tokens
RTX 3060 · 12 GB VRAMFast memory holds itQ8_0GGUF · 3.3 GB · needs ~7 GB65,536 tokens~5 GBFast memory holds itQ8_0GGUF · 3.3 GB · needs ~7 GB65,536 tokens
RTX 5080 · 16 GB VRAMFast memory holds itQ8_0GGUF · 3.3 GB · needs ~7 GB65,536 tokens~5 GBFast memory holds itQ8_0GGUF · 3.3 GB · needs ~7 GB65,536 tokens
RTX 4090 · 24 GB VRAMFast memory holds itQ8_0GGUF · 3.3 GB · needs ~7 GB65,536 tokens~5 GBFast memory holds itQ8_0GGUF · 3.3 GB · needs ~7 GB65,536 tokens
RTX 5090 · 32 GB VRAMFast memory holds itQ8_0GGUF · 3.3 GB · needs ~7 GB65,536 tokens~5 GBFast memory holds itQ8_0GGUF · 3.3 GB · needs ~7 GB65,536 tokens
DGX H100 · 8 × 80 GBFast memory holds itQ8_0GGUF · 3.3 GB · needs ~21 GB65,536 tokens~19 GBFast memory holds itQ8_0GGUF · 3.3 GB · needs ~21 GB65,536 tokens

The file column uses the site's quant ladder (Q8_0 to Q3_K_M, MLX 8 to 4-bit on a Mac): the largest collected file whose size × 1.2 plus the context cache and the machine's reserve fits the fast memory. Planning estimates, not measured runs.

Plan another quantization or context

Consider it for

Compare SmolLM3 3B using its publisher card. No overall score appears in the LiveBench 2026-06-25 snapshot.

Check first

The Q4 weights plus working room come to about 3 GB from 3B stated parameters, before the machine's reserve (Runs on adds it per machine). It is not a measured minimum; context, batch size, runtime and quantization change actual memory use.

Versions to download

Files you can download for this checkpoint. The section opens on the most downloaded collected GGUF source, bartowski, at Q4_K_M. On a Mac it opens on the most downloaded MLX source. Sizes are the Hub's published bytes. What is Q4?

bartowski

bartowski/HuggingFaceTB_SmolLM3-3B-GGUF · HuggingFaceTB_SmolLM3-3B-Q4_K_M.gguf · 1.9 GB

Run this file

ollama run hf.co/bartowski/HuggingFaceTB_SmolLM3-3B-GGUF:Q4_K_M
llama-server -hf bartowski/HuggingFaceTB_SmolLM3-3B-GGUF:Q4_K_M
lms get https://huggingface.co/bartowski/HuggingFaceTB_SmolLM3-3B-GGUF
  • mlx-lm: mlx-lm runs MLX repositories; pick the MLX version on a Mac

Quality scores on this page belong to the original checkpoint; these files are not scored or ranked. File sizes read 2026-10-23 · download counts read 2026-10-17 · commands: Hugging Face Hub docs: Use Ollama with any GGUF model · llama.cpp server README (-hf <user>/<model>[:quant]) · LM Studio docs: lms get with a Hugging Face URL · mlx-lm README (mlx_lm.generate --model)

Fine-tunes, adapters and other versions169 fine-tunes
Quantized versions
125
Fine-tunes
169
Adapters · merges
63 · 3

Browse them on the Hugging Face model tree · read 2026-10-23

Uncensored and abliterated versions7

Some community uploads remove the model's safety refusals. Quality may drop, the original publisher does not support them, claims about them are the uploader's, and the original license still applies. This site never recommends or ranks them.

7 derivative repositories have “uncensored” or “abliterated” in their name (read 2026-10-23).

Measured speed

No published measurement for this checkpoint

Tokens per second are shown only when a source publishes them with hardware, build and setup. They are never estimated.

How to run

The memory needed above is the Q4 load before a machine's reserve. Formula: parameters × 0.625 bytes, plus 20% working allowance; each machine then adds its system reserve: 25% of unified or system memory, 2 to 8 GB, or 2 GB per GPU on dedicated cards. This is neither measured VRAM nor checkpoint download size.

Q4 planning estimate only. Check checkpoint format, runtime support, KV cache, context, batch size and offload before choosing hardware.

Evidence level: Generic estimate only; no execution reproduced by this site.

Publisher sampling defaults
temperature 0.6 · top_p 0.95generation_config.json · read 2026-10-23source
Chat template markers
tool calls · thinkingmarkers found in the template, not a test of the feature · read 2026-10-23source
Before downloading
  1. Check the exact checkpoint and license linked here.
  2. Confirm a compatible runtime and precision or quantization for your platform.
  3. Allow for context, batch size and any extra media components beyond the weights.

Technical details

What the source and the catalog store for this checkpoint. Missing fields are left out, never zero.

Checkpoint / version
HuggingFaceTB/SmolLM3-3B
Task and modality
language and reasoning · text
Count and basis
3B · manufacturer specificationHuggingFaceTB/SmolLM3-3B model card (Hugging Face) · read 2026-10-17source
Context
65,536 tokens published for this modelHuggingFaceTB/SmolLM3-3B model card (Hugging Face) · read 2026-10-17source
Architecture
DenseHuggingFaceTB/SmolLM3-3B model card (Hugging Face) · read 2026-10-17source
Released
Hugging Face blog index (huggingface/blog _blog.yml, smollm3 entry) · read 2026-10-17source
Runtime and precision
No local execution verified.
Languages
en, fr, es, it, pt, zh, ar, ruModel card metadata · read 2026-10-25source
Commercial use
Allowed Apache License 2.0 text (SPDX) · read 2026-10-18license text
Every stored field12 fields
Revision checked
a07cc9a04f16550a088caea529712d1d335b0ac1
Architecture
SmolLM3ForCausalLM
Library
transformers
Hub task tag
text-generation
Declared base models
HuggingFaceTB/SmolLM3-3B-Base
Formats
safetensors, pt
Access
Open download
Publisher kind
Official
Repository created
2025-07-08creation is not the model's release
Repository updated
2025-09-10a commit is not a release date
Hub downloads (30 days) · likes
718,665 · 1,070this repository only · read 2026-10-25
Catalog state
Hub config or weights found; no inventory benchmark link

Sources

Identity and published specifications come from the checkpoint page. Benchmark scores belong to the named release and do not measure speed on your computer.

Catalog reviewed: 2026-10-17. Benchmark release: 2026-06-25; snapshot collected: 2026-10-18.