Hugging Face Hub Models
The Hugging Face Hub hosts 500K+ pre-trained models. Here are the most useful free models by category.
How to find models
- Browse by task: huggingface.co/models — filter by task, library, language
- Trending: huggingface.co/models?sort=trending
- ONNX-ready: huggingface.co/models?library=onnx — can run in browser
Top Free Models by Task
Text Generation (LLMs)
| Model |
Parameters |
License |
Notes |
Link |
| Llama 3.1 8B |
8B |
Llama 3.1 |
Meta's latest, instruction-tuned |
HF |
| Mistral 7B |
7B |
Apache-2.0 |
Fast, high quality |
HF |
| Gemma 2 9B |
9B |
Gemma |
Google's open model |
HF |
| Phi-3 Mini |
3.8B |
MIT |
Microsoft, small but capable |
HF |
| Qwen2.5 7B |
7B |
Apache-2.0 |
Strong multilingual |
HF |
| TinyLlama |
1.1B |
Apache-2.0 |
Smallest useful LLM |
HF |
Text Embeddings
| Model |
Dimensions |
License |
Best For |
Link |
| all-MiniLM-L6-v2 |
384 |
Apache-2.0 |
Fast semantic search |
HF |
| bge-small-en-v1.5 |
384 |
MIT |
RAG / retrieval |
HF |
| e5-large-v2 |
1024 |
MIT |
High quality embeddings |
HF |
| gte-base |
768 |
MIT |
General purpose |
HF |
| nomic-embed-text-v1.5 |
768 |
Apache-2.0 |
Long context (8K) |
HF |
Image Generation
| Model |
Resolution |
License |
Notes |
Link |
| Stable Diffusion XL |
1024x1024 |
CreativeML Open |
Best open image gen |
HF |
| Stable Diffusion 3 |
1024x1024 |
Stability Community |
Latest SD |
HF |
| FLUX.1 Schnell |
1024x1024 |
Apache-2.0 |
Fast generation |
HF |
| Kandinsky 3 |
1024x1024 |
Apache-2.0 |
Multilingual prompts |
HF |
Speech & Audio
| Model |
Task |
License |
Notes |
Link |
| Whisper Large v3 |
Speech-to-text |
MIT |
Best open ASR, 100 languages |
HF |
| Bark |
Text-to-speech |
MIT |
Multilingual, sound effects |
HF |
| MusicGen |
Music generation |
CC-BY-NC |
Text-to-music |
HF |
| Wav2Vec2 |
Speech recognition |
Apache-2.0 |
Fine-tunable |
HF |
Object Detection
| Model |
Objects |
License |
Speed |
Link |
| YOLOv8 |
80 COCO classes |
AGPL-3.0 |
Very fast |
HF |
| DETR |
91 COCO classes |
Apache-2.0 |
Medium |
HF |
| OWL-ViT |
Open vocabulary |
Apache-2.0 |
Slow |
HF |
| Grounding DINO |
Open vocabulary |
Apache-2.0 |
Medium |
HF |
Segmentation
| Model |
Type |
License |
Link |
| SAM |
Segment Anything |
Apache-2.0 |
HF |
| Mask2Former |
Panoptic segmentation |
MIT |
HF |
| SegFormer |
Semantic segmentation |
Apache-2.0 |
HF |