Explore public Hugging Face models by task, license, size and estimated VRAM. Compare model metadata before you commit engineering time or GPU budget.
sentence-transformers/all-MiniLM-L6-v2 is a sentence similarity model indexed for deployment research. Estimated minimum GPU memory is 4 GB. It is publicly listed on Hugging Face.
cross-encoder/ms-marco-MiniLM-L6-v2 is a text ranking model indexed for deployment research. Estimated minimum GPU memory is 8 GB. It is publicly listed on Hugging Face.
BAAI/bge-small-en-v1.5 is a feature extraction model indexed for deployment research. Estimated minimum GPU memory is 4 GB. It is publicly listed on Hugging Face.
argmaxinc/whisperkit-coreml is a automatic speech recognition model indexed for deployment research. Estimated minimum GPU memory is 8 GB. It is publicly listed on Hugging Face.
Comfy-Org/MiniMax-H3 is a AI workload model indexed for deployment research. Estimated minimum GPU memory is 8 GB. It is publicly listed on Hugging Face.
openai-community/gpt2 is a text generation model indexed for deployment research. Estimated minimum GPU memory is 8 GB. It is publicly listed on Hugging Face.
google-bert/bert-base-uncased is a fill mask model indexed for deployment research. Estimated minimum GPU memory is 8 GB. It is publicly listed on Hugging Face.
BAAI/bge-m3 is a sentence similarity model indexed for deployment research. Estimated minimum GPU memory is 4 GB. It is publicly listed on Hugging Face.
sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2 is a sentence similarity model indexed for deployment research. Estimated minimum GPU memory is 4 GB. It is publicly listed on Hugging Face.
autogluon/chronos-2 is a time series forecasting model indexed for deployment research. Estimated minimum GPU memory is 8 GB. It is publicly listed on Hugging Face.
google-t5/t5-small is a translation model indexed for deployment research. Estimated minimum GPU memory is 8 GB. It is publicly listed on Hugging Face.
Qwen/Qwen3-0.6B is a text generation model indexed for deployment research. Estimated minimum GPU memory is 4 GB. It is publicly listed on Hugging Face.
Find candidates by task, license, parameters and real-world hardware constraints.
Put up to four models side by side before downloading weights or paying for compute.
Use transparent memory estimates as a starting point, then verify at the source.
AI Pentium organizes public metadata from the Hugging Face Hub API, including task, library, downloads, likes and declared license.
Parameter counts and model signals are converted into directional GPU VRAM tiers so you can shortlist models faster.
Each result links to the original Hugging Face repository where you should confirm weights, model-card guidance, license terms and limitations.
Featured listings, launch packages and independent model evaluations are available.