File size: 12,377 Bytes
d2c5391 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 |
---
base_model:
- HuggingFaceTB/SmolLM3-3B-Base
library_name: transformers
license: apache-2.0
language:
- en
- fr
- es
- it
- pt
- zh
- ar
- ru
tags:
- transformers.js
- unsloth
---
<div>
<p style="margin-top: 0;margin-bottom: 0;">
<em><a href="https://docs.unsloth.ai/basics/unsloth-dynamic-v2.0-gguf">Unsloth Dynamic 2.0</a> achieves superior accuracy & outperforms other leading quants.</em>
</p>
<div style="display: flex; gap: 5px; align-items: center; ">
<a href="https://github.com/unslothai/unsloth/">
<img src="https://github.com/unslothai/unsloth/raw/main/images/unsloth%20new%20logo.png" width="133">
</a>
<a href="https://discord.gg/unsloth">
<img src="https://github.com/unslothai/unsloth/raw/main/images/Discord%20button.png" width="173">
</a>
<a href="https://docs.unsloth.ai/">
<img src="https://raw.githubusercontent.com/unslothai/unsloth/refs/heads/main/images/documentation%20green%20button.png" width="143">
</a>
</div>
</div>
# SmolLM3
## Table of Contents
1. [Model Summary](#model-summary)
2. [How to use](#how-to-use)
3. [Evaluation](#evaluation)
4. [Training](#training)
5. [Limitations](#limitations)
6. [License](#license)
## Model Summary
SmolLM3 is a 3B parameter language model designed to push the boundaries of small models. It supports 6 languages, advanced reasoning and long context. SmolLM3 is a fully open model that offers strong performance at the 3B–4B scale.
**SmolLM3-3B-Base** is the base model after pretraining, you can find the instruct model at [SmolLM3-3B](https://huggingface.co/HuggingFaceTB/SmolLM3-3B).
The model is a decoder-only transformer using GQA and NoPE, it was pretrained on 11.2T tokens with a staged curriculum of web, code, math and reasoning data. Post-training included midtraining on 140B reasoning tokens followed by supervised fine-tuning and alignment via Anchored Preference Optimization (APO).
### Key features
- Instruct model optimized for **hybrid reasoning**
- **Fully open model**: open weights + full training details including public data mixture and training configs
- **Long context:** Trained on 64k context and suppots up to **128k tokens** using YARN extrapolation
- **Multilingual**: 6 natively supported (English, French, Spanish, German, Italian, and Portuguese)
For more details refer to our blog post: https://hf.co/blog/smollm3
### How to use
The modeling code for SmolLM3 is available in transformers `v4.53.0`, so make sure to upgrade your transformers version. You can also load the model with the latest `vllm` which uses transformers as a backend.
```bash
pip install -U transformers
```
```python
from transformers import AutoModelForCausalLM, AutoTokenizer
checkpoint = "HuggingFaceTB/SmolLM3-3B"
device = "cuda" # for GPU usage or "cpu" for CPU usage
tokenizer = AutoTokenizer.from_pretrained(checkpoint)
# for multiple GPUs install accelerate and do `model = AutoModelForCausalLM.from_pretrained(checkpoint, device_map="auto")`
model = AutoModelForCausalLM.from_pretrained(checkpoint).to(device)
inputs = tokenizer.encode("Gravity is", return_tensors="pt").to(device)
outputs = model.generate(inputs)
print(tokenizer.decode(outputs[0]))
```
For local inference, you can use `llama.cpp`, `ONNX`, `MLX` and `MLC`. You can find quantized checkpoints in this collection (https://huggingface.co/collections/HuggingFaceTB/smollm3-686d33c1fdffe8e635317e23).
### Long context processing
The current `config.json` is set for context length up to 65,536 tokens. To handle longer inputs (128k or 256k), we utilize YaRN you can change the `max_position_embeddings` and rope_scaling` to:
```
{
...,
"rope_scaling": {
"factor": 2.0, #2x65536=131 072
"original_max_position_embeddings": 65536,
"type": "yarn"
}
}
```
## Evaluation
In this section, we report the evaluation results of SmolLM3 model. All evaluations are zero-shot unless stated otherwise, and we use [lighteval](https://github.com/huggingface/lighteval) to run them.
We highlight the best score in bold and underline the second-best score.
### Base Pre-Trained Model
#### English benchmarks
Note: All evaluations are zero-shot unless stated otherwise. For Ruler 64k evaluation, we apply YaRN to the Qwen models with 32k context to extrapolate the context length.
| Category | Metric | SmolLM3-3B | Qwen2.5-3B | Llama3-3.2B | Qwen3-1.7B-Base | Qwen3-4B-Base |
|---------|--------|---------------------|------------|--------------|------------------|---------------|
| Reasoning & Commonsense| HellaSwag | **76.15** | 74.19 |<u>75.52</u> | 60.52 | 74.37 |
| | ARC-CF (Average) | **65.61** | 59.81 | 58.58 | 55.88 | <u>62.11</u> |
| | Winogrande | 58.88 | **61.41** | 58.72 | 57.06 | <u>59.59</u> |
| | CommonsenseQA | <u>55.28</u> | 49.14 | **60.60** | 48.98 | 52.99 |
| Knowledge & Understanding | MMLU-CF (Average) | <u>44.13</u> | 42.93 | 41.32 | 39.11 | **47.65** |
| | MMLU Pro CF | <u>19.61</u> | 16.66 | 16.42 | 18.04 | **24.92** |
| | MMLU Pro MCF | <u>32.70</u> | 31.32 | 25.07 | 30.39 | **41.07** |
| | PIQA | **78.89** | 78.35 | <u>78.51</u> | 75.35 | 77.58 |
| | OpenBookQA | 40.60 | 40.20 | <u>42.00</u> | 36.40 | **42.40** |
| | BoolQ | **78.99** | 73.61 | <u>75.33</u> | 74.46 | 74.28 |
| **Math & Code** | | | | | | |
| Coding & math | HumanEval+ | 30.48 | 34.14| 25.00 | <u>43.29</u>| **54.87** |
| | MBPP+ | 52.91 | 52.11 | 38.88| <u>59.25</u> | **63.75** |
| | MATH (4-shot) | <u>46.10</u> | 40.10 | 7.44 | 41.64 | **51.20** |
| | GSM8k (5-shot) | 67.63 | <u>70.13</u> | 25.92 | 65.88 | **74.14** |
| **Long context** | | | | | | |
| | Ruler 32k | 76.35 | 75.93 | <u>77.58</u> | 70.63 | **83.98** |
| | Ruler 64k | <u>67.85</u> | 64.90 | **72.93** | 57.18 | 60.29 |
| | Ruler 128k | 61.03 | <u>62.23</u> | **71.30** | 43.03 | 47.23 |
#### Multilingual benchmarks
| Category | Metric | SmolLM3 3B Base | Qwen2.5-3B | Llama3.2 3B | Qwen3 1.7B Base | Qwen3 4B Base |
|---------|--------|---------------------|------------|--------------|------------------|---------------|
| Main supported languages | | | | | | | |
| French| MLMM Hellaswag | **63.94** | 57.47 | 57.66 | 51.26 | <u>61.00</u> |
| | Belebele | 51.00 | <u>51.55</u> | 49.22 |49.44| **55.00** |
| | Global MMLU (CF) | <u>38.37</u> | 34.22 | 33.71 | 34.94 |**41.80** |
| | Flores-200 (5-shot) | 62.85| 61.38| <u>62.89<u/u> | 58.68 | **65.76** |
| Spanish| MLMM Hellaswag | **65.85** | 58.25 | 59.39 | 52.40 | <u>61.85</u> |
| | Belebele | 47.00 | <u>48.88</u> | 47.00 | 47.56 | **50.33** |
| | Global MMLU (CF) | <u>38.51</u> | 35.84 | 35.60 | 34.79 |**41.22** |
| | Flores-200 (5-shot) | <u>48.25</u>| 50.00| 44.45 | 46.93 | **50.16** |
| German| MLMM Hellaswag | **59.56** | 49.99| 53.19|46.10| <u>56.43</u>|
| | Belebele | <u>48.44</u> | 47.88 | 46.22 | 48.00 | **53.44**|
| | Global MMLU (CF) | <u>35.10</u> | 33.19 | 32.60 | 32.73 |**38.70** |
| | Flores-200 (5-shot) | **56.60**| 50.63| <u>54.95</u> | 52.58 | 50.48 |
| Italian| MLMM Hellaswag | **62.49** | 53.21 | 54.96 | 48.72 | <u>58.76</u> |
| | Belebele | <u>46.44</u> | 44.77 | 43.88 | 44.00 | **48.78** | 44.88 |
| | Global MMLU (CF) | <u>36.99</u> | 33.91 | 32.79 | 35.37 |**39.26** |
| | Flores-200 (5-shot) | <u>52.65<u/>| **54.87**| 48.83 | 48.37 | 49.11 |
| Portuguese| MLMM Hellaswag | **63.22** | 57.38 | 56.84 | 50.73 | <u>59.89</u> |
| | Belebele | 47.67 | **49.22** | 45.00 | 44.00 | 50.00 | <u>49.00</U> |
| | Global MMLU (CF) | <u>36.88</u> | 34.72 | 33.05 | 35.26 |**40.66** |
| | Flores-200 (5-shot) | <u>60.93</u> |57.68| 54.28 | 56.58 | **63.43** |
The model has also been trained on Arabic (standard), Chinese and Russian data, but has seen fewer tokens in these languages compared to the 6 above. We report the performance on these langages for information.
| Category | Metric | SmolLM3 3B Base | Qwen2.5-3B | Llama3.2 3B | Qwen3 1.7B Base | Qwen3 4B Base |
|---------|--------|---------------------|------------|--------------|------------------|---------------|
| Other supported languages | | | | | | | |
| Arabic| Belebele | 40.22 | 44.22 | <u>45.33</u> | 42.33 | **51.78** |
| | Global MMLU (CF) | 28.57 | 28.81 | 27.67 | <u>29.37</u> | **31.85** |
| | Flores-200 (5-shot) | <u>40.22</u> | 39.44 | **44.43** | 35.82 | 39.76 |
| Chinese| Belebele | 43.78 | 44.56 | <u>49.56</u> | 48.78 | **53.22** |
| | Global MMLU (CF) | 36.16 | 33.79 | <u>39.57</u> | 38.56 | **44.55** |
| | Flores-200 (5-shot) | 29.17 | **33.21** | 31.89 | 25.70 | <u>32.50</u> |
| Russian| Belebele | <u>47.44</u> | 45.89 | <u>47.44</u> | 45.22 | **51.44** |
| | Global MMLU (CF) | <u>36.51</u> | 32.47 | 34.52 | 34.83 | **38.80** |
| | Flores-200 (5-shot) | 47.13 | 48.74 | 50.74 | <u>54.70</u> | **60.53** |
### Instruction Model
#### No Extended Thinking
Evaluation results of non reasoning models and reasoning models in no thinking mode. We highlight the best and second-best scores in bold.
| Category | Metric | SmoLLM3-3B | Qwen2.5-3B | Llama3.1-3B | Qwen3-1.7B | Qwen3-4B |
|---------|--------|------------|------------|-------------|------------|----------|
| High school math competition | AIME 2025 | <u>9.3</u> | 2.9 | 0.3 | 8.0 | **17.1** |
| Math problem-solving | GSM-Plus | 72.8 | <u>74.1</u> | 59.2 | 68.3 | **82.1** |
| Competitive programming | LiveCodeBench v4 | <u>15.2</u> | 10.5 | 3.4 | 15.0 | **24.9** |
| Graduate-level reasoning | GPQA Diamond | <u>35.7</u> | 32.2 | 29.4 | 31.8 | **44.4** |
| Instruction following | IFEval | **76.7** | 65.6 | 71.6 | <u>74.0</u> | 68.9 |
| Alignment | MixEval Hard | 26.9 | <u>27.6</u> | 24.9 | 24.3 | **31.6** |
| Tool Calling | BFCL| <u>92.3</u> | - | <u>92.3</u> * | 89.5 | **95.0** |
| Multilingual Q&A | Global MMLU | <u>53.5</u> | 50.54 | 46.8 | 49.5 | **65.1** |
(*): this is a tool calling finetune
#### Extended Thinking
Evaluation results in reasoning mode for SmolLM3 and Qwen3 models:
| Category | Metric | SmoLLM3-3B | Qwen3-1.7B | Qwen3-4B |
|---------|--------|------------|------------|----------|
| High school math competition | AIME 2025 | <u>36.7</u> | 30.7 | **58.8** |
| Math problem-solving | GSM-Plus | <u>83.4</u> | 79.4 | **88.2** |
| Competitive programming | LiveCodeBench v4 | 30.0 | <u>34.4</u> | **52.9** |
| Graduate-level reasoning | GPQA Diamond | <u>41.7</u> | 39.9 | **55.3** |
| Instruction following | IFEval | 71.2 | <u>74.2</u> | **85.4** |
| Alignment | MixEval Hard | 30.8 | <u>33.9</u> | **38.0** |
| Tool Calling | BFCL | <u>88.8</u> | <u>88.8</u> | **95.5** |
| Multilingual Q&A | Global MMLU | <u>64.1</u> | 62.3 | **73.3** |
## Training
### Model
- **Architecture:** Transformer decoder
- **Pretraining tokens:** 11T
- **Precision:** bfloat16
### Software & hardware
- **GPUs:** 384 H100
- **Training Framework:** [nanotron](https://github.com/huggingface/nanotron/tree/main)
- **Data processing framework:** [datatrove](https://github.com/huggingface/datatrove)
- **Evaluation framework:** [lighteval](https://github.com/huggingface/lighteval)
- **Post-training Framework:** [TRL](https://github.com/huggingface/trl)
### Open resources
Here is an infographic with all the training details.
- The datasets used for pretraining can be found in this [collection](https://huggingface.co/collections/HuggingFaceTB/smollm3-pretraining-datasets-685a7353fdc01aecde51b1d9) and those used in mid-training and post-training will be released in the following weeks
- The training and evaluation configs and code can be found in the [huggingface/smollm](https://github.com/huggingface/smollm) repository.
## Limitations
SmolLM3 can produce text on a variety of topics, but the generated content may not always be factually accurate, logically consistent, or free from biases present in the training data. These models should be used as assistive tools rather than definitive sources of information. Users should always verify important information and critically evaluate any generated content.
## License
[Apache 2.0](https://www.apache.org/licenses/LICENSE-2.0) |