Leaderboard / Hugging Face (SmolLM)

SmolLM3 3B

Name: SmolLM3 3B
Author: Hugging Face (SmolLM)

Hugging Face (SmolLM) · SmolLM · released 2025-07-08 · data checked 2026-11-01

SmolLM3 3B is a 3.1B-parameter dense model from Hugging Face (SmolLM) (Europe), released 2025-07-08 under the Apache-2.0 license. It scores 30.4 on the Small LM Index, #15 of 21 tiny models. It needs about 3 GB of memory at 4-bit with an 8K-token context.

Small LM Index

30.4

#15 of 21 scored tiny models

Likely 26.5–34.7 · from 3 of 6 core benchmarks

Parameters

3.1B

Dense

Price per 1M tokens

—

No hosted API on OpenRouter; self-host

Context window

64K

tokens

Benchmarks

Each bar shows this model's score. The tick marks the best tiny model on that benchmark. Sources: 5 reported by Hugging Face (SmolLM). How we label sources.

GPQA Diamond Lab-reported Knowledge41.7

Class best: Nanbeige4.2 3B 87.4

AIME Lab-reported Math · AIME 2025, thinking mode36.7

Class best: Spark-X2.6 4B 91.2

LiveCodeBench Lab-reported Coding · LiveCodeBench v430.0

Class best: LFM3 1.7B 73.5

Multilingual MMLU Lab-reported Multilingual · Global-MMLU64.1

IFEval Lab-reported Instruction following71.2

Class best: LFM3 1.7B 96.7

Not reported: estimated

What the index model expects from this model's other scores. Not a measurement.

MMLU-Pro≈ 58

Humanity's Last Exam≈ 2

SWE-bench Verified≈ 12

Specifications

Lab: Hugging Face (SmolLM) (FR · Europe)
Architecture: Dense transformer
Total parameters: 3.1B
Active parameters: All (dense)
Size class: Tiny
Context window: 64K tokens
Input modalities: text
Reasoning: Hybrid (thinking on/off)
License: Apache-2.0
Openness: Fully open: weights, training data and code
Memory to run: ~3 GB at 4-bit, 8K context (more sizes)
Hugging Face: 567.6K downloads (30 days) · 1.1K likes

Weights on Hugging Face ↗

Compare with

SmolLM3 3B vs Granite 4.2 3B (3B, index 54.8)
SmolLM3 3B vs Ministral 3 3B (3B, index 49.4)
SmolLM3 3B vs LFM2.5 2.6B (2.7B, index 43.2)
SmolLM3 3B vs MiniCPM5 2B (2.5B, index 57.5)

Ranked in

Sources

https://huggingface.co/HuggingFaceTB/SmolLM3-3B

Something wrong here? Report it.

Run it locally

Uses ggml-org/SmolLM3-3B-GGUF (Q4_K_M). Swap the tag for another size: F16, Q4_K_M, Q8_0.

Ollama

ollama run hf.co/ggml-org/SmolLM3-3B-GGUF:Q4_K_M

llama.cpp

llama-server -hf ggml-org/SmolLM3-3B-GGUF:Q4_K_M

Then open http://localhost:8080 for a chat window.

MLX (Mac)

pip install -U mlx-lm
mlx_lm.chat --model mlx-community/SmolLM3-3B-4bit

vLLM (GPU server)

vllm serve HuggingFaceTB/SmolLM3-3B

Full-precision weights, for a server GPU.

Open in LM Studio

Memory needed: weights, context cache and runtime buffers.

Quantization	4K context	32K context	64K context
16-bit F16	7.0 GB	9.1 GB	11 GB
8-bit Q8_0	4.1 GB	6.3 GB	8.6 GB
4-bit Q4_K_M	2.7 GB	4.8 GB	7.1 GB

Check it against your GPU, Mac or phone →