Project AURIONAI Use-Cases Research & Innovation
Hugging Face
edge LLM

SmolLM3-3B

Hugging Face's 3B small model, historically fully open; no SmolLM4 found.

image/png.
Levelmodel
Classedge LLM
Parameters3B
Context128,000
Released8 Jul 2025
Opennessfully open
Runs onlaptop, mobile or NPU, CPU server
Statusactive

From the model's own page; who made each picture and where it comes from are listed on the Image Credits page.

Benchmark Results

vendor-reported
Card 3-model comparison table; mode label not extractedvendor-reported
36.7
050100

Score in %. Higher is better.

Card 5-model comparison table; mode label not extractedvendor-reported
9.3
050100

Score in %. Higher is better.

BFCLagents and tool use
Card 5-model comparison table; mode label not extractedvendor-reported
92.3
050100

Score in %. Higher is better.

Global MMLUknowledge and reasoning
Card 5-model comparison table; mode label not extractedvendor-reported
53.5
050100

Score in %. Higher is better.

GPQA Diamondknowledge and reasoning
Card 5-model comparison table; mode label not extractedvendor-reported
35.7
050100

Score in %. Higher is better.

IFEvalknowledge and reasoning
Card 5-model comparison table; mode label not extractedvendor-reported
76.7
050100

Score in %. Higher is better.

Where this comes from
Every score names who ran it: vendor-reported, maintainer-run or independent. A rank is shown with the population it was ranked in (gap rule 3).

Lineage

instruct of
SmolLM3-3B-Base to SmolLM3-3B: instruct of, high confidenceSmolLM3-3B-BaseSmolLM3-3B-BaseHugging FaceSmolLM3-3BSmolLM3-3BHugging Face, Jul 2025
Hugging FaceSmolLM3-3B-Base is the parent (instruct of)high confidenceHugging Face base_model
Where this comes from
Parent and child relations from model cards, Hugging Face base_model fields and the genealogy tables of the research notes. The tree shows two generations each way, 2 models. Confidence and evidence are the notes' own.

R&D Topics

Identity

Developer
Hugging Face
Family
not in the notes
Released
8 Jul 2025 sourcesnotes
Weights published
–
Record checked
10 Oct 2026
Where this comes from
Developer and dates from the research notes. Release date is the vendor's public announcement; weights date is when open weights were published.

Architecture and Size

Architecture
dense transformer Decoder-only with GQA and NoPE (3:1); dual-mode reasoning (/think, /no_think); pretrained from scratch on 11.2T tokens, then 140B reasoning tokens of midtraining, SFT and APO; trained on 384 H100
Total parameters
3B sources
Active parameters
–
Checkpoint parameters
–
Note
Trained at 64K context, 128K via YaRN
Context length
128,000 tokens
Input
text
Output
text
Where this comes from
Parameter counts as the developer states them. Where the checkpoint holds a different total, both are shown (register A16).

Openness and Access

Openness
fully open
Licence
Apache License 2.0 sources
Access routes
open weights download
Where this comes from
The licence file governs over registry tags and aggregator labels.

Deployment

Runs on
laptop, mobile or NPU, CPU server
Footprint
Quantized footprint unverified
Runtimes
not in the notes
Where this comes from
Deployment tiers and runtimes as named in the notes. Footprint is the notes' figure, not a measurement by AURION.

Indonesia

Indonesian support
not stated
Regional languages
not in the notes
Jakarta hosting
unknown
Languages
6 languages
Where this comes from
Indonesia fields are carried on every model; unknown is an allowed value.

Industry Use Cases

No industry use case links this model.

Sources

  1. HuggingFaceTB (Hugging Face Smol Models Research) huggingface.co https://huggingface.co/HuggingFaceTB
  2. HuggingFaceTB/SmolLM3-3B , Hugging Face huggingface.co https://huggingface.co/HuggingFaceTB/SmolLM3-3B
  3. huggingface.co/api/models/HuggingFaceTB/SmolLM3-3B https://huggingface.co/api/models/HuggingFaceTB/SmolLM3-3B
  4. SmolLM by HuggingFace | European Open Source AI Index osai-index.eu https://osai-index.eu/model/smollm
  5. SmolLM3: smol, multilingual, long-context reasoner huggingface.co https://huggingface.co/blog/smollm3
  6. huggingface.co/datasets/HuggingFaceTB/smoltalk2 https://huggingface.co/datasets/HuggingFaceTB/smoltalk2

To cite this page: Project AURION, AI Use-Cases Research & Innovation, "SmolLM3-3B", https://aurion.semai.id/models/hugging-face-smollm3-3b/, records of October 2026.