AImodelsComparison.com 1,222 open · 200 API
Menu
Models Best of Trending Can I run it? Methodology About
1,222 models · 1,222 open weights · 200 API-only · refreshed nightly

Open-weights models

Models whose weights are published on Hugging Face — download them and run them on your own hardware. Many are also served by hosted providers; those carry an "Also via API" badge with live prices.

Type
All 1,422 Open weights (self-host) 1,222 API-only 200
Task
All text-generation 533 text-to-image 155 image-text-to-text 110 ASR 87 sentence-similarity 65 translation 55 text-classification 51 feature-extraction 48 image-classification 39 token-classification 21 text-to-speech 17 audio-classification 12 image-segmentation 10 object-detection 9 summarization 8 audio-text-to-text 1 any-to-any 1
Runs on
Any RTX 3060 · 12 GB RTX 4070 · 16 GB RTX 4090 · 24 GB 2 × 24 GB Mac M3 · 24 GB Mac M4 Max · 64 GB
Licence
Size
$
1,222 models
#ModelParamsContextCommercial use30dMin VRAM
751 TwIL-LM3 webAI-Official · text-generation Open weights 3.1 B 66 K other 242 K from 3.1 GB
752 h2ovl-mississippi-800m h2oai · text-generation Open weights 830 M — ✓ apache-2.0 242 K from 2.4 GB
753 DreamShaper Lykon · text-to-image Open weights — — other 241 K from 185.3 GB
754 convnext_femto.d1_in1k timm · image-classification Open weights 10 M — ✓ apache-2.0 241 K from 0.5 GB
755 Qwen3-8B-FP8 nvidia · text-generation Open weights 8.2 B 41 K ✓ apache-2.0 240 K from 12.1 GB
756 mit-b2 nvidia · image-classification Open weights — — other 240 K —
757 CommunityForensics-DeepfakeDet-ViT buildborderless · image-classification Open weights 20 M — ✓ mit 240 K from 0.9 GB
758 Vikhr-Nemo-12B-Instruct-R-21-09-24 Vikhrmodels · text-generation Open weights 12.3 B 1.0 M ✓ apache-2.0 239 K from 30.8 GB
759 POCKET-Darwin-180B-GGUF FINAL-Bench · text-generation Open weights 180.0 B — other 239 K from 82.3 GB
760 gemma-3-270m google · text-generation · gated Open weights 270 M — ⚠ gemma 238 K from 1.1 GB
761 s1-mini-GGUF superwhisper · text-generation Open weights — — other 237 K from 1.0 GB
762 Mistral-Small-3.1-24B-Instruct-2503 mistralai · text-generation Open weights Also via API 24.0 B — ✓ apache-2.0 237 K from 109.8 GB
763 Qwen3-1.7B PrimeIntellect · text-generation Open weights 2.0 B 41 K ✓ apache-2.0 236 K from 5.3 GB
764 indictrans2-en-indic-dist-200M ai4bharat · translation · gated Open weights 275 M — ✓ mit 234 K from 1.7 GB
765 POCKET-26B-GGUF FINAL-Bench · text-generation Open weights 26.0 B — ✓ apache-2.0 233 K from 16.6 GB
766 diffusiongemma-26B-A4B-it-NVFP4 nvidia · text-generation Open weights 14.4 B — ✓ apache-2.0 232 K from 23.4 GB
767 GLM-4.7-Flash-GGUF unsloth · text-generation Open weights — — ✓ mit 232 K from 13.0 GB
768 Qwen2.5-Math-1.5B Qwen · text-generation Open weights 1.5 B 4 K ✓ apache-2.0 231 K from 4.1 GB
769 gpt-oss-20b-GGUF ggml-org · text-generation Open weights 20.0 B 131 K ✓ apache-2.0 231 K from 4.5 GB
770 h2ovl-mississippi-2b h2oai · text-generation Open weights 2.2 B — ✓ apache-2.0 230 K from 5.6 GB
771 IP-Adapter-FaceID h94 · text-to-image Open weights — — unknown 230 K from 1.5 GB
772 Kimi-K2.5-NVFP4 nvidia · text-generation Open weights — — other 229 K from 650.4 GB
773 Kimi-K2-Thinking-NVFP4 nvidia · text-generation Open weights 519.1 B 262 K other 229 K from 732.0 GB
774 Llama-3.1-8B-Instruct-4bit mlx-community · text-generation Open weights 1.3 B 131 K ⚠ llama3.1 228 K from 5.7 GB
775 Llama-3.2-3B meta-llama · text-generation · gated Open weights 3.2 B — ⚠ llama3.2 228 K from 8.0 GB
776 Spark-X2.5-1.7B-GGUF XHToken · text-generation Open weights 1.7 B — ✓ apache-2.0 227 K from 2.0 GB
777 vlt5-base-keywords Voicelab · text-generation Open weights 280 M — ✓ cc-by-4.0 227 K from 1.8 GB
778 BiRefNet_lite ZhengPeng7 · image-segmentation Open weights 44 M — ✓ mit 226 K from 0.7 GB
779 distilbert-multilingual-toxicity-classifier-ONNX onnx-community · text-classification Open weights — 512 ⚠ openrail++ 225 K —
780 Ornith-1.5-397B ornith-ai · text-generation Open weights 403.4 B — ✓ mit 225 K from 948.5 GB
781 Qwen3-1.7B-GGUF unsloth · text-generation Open weights 1.7 B 41 K ✓ apache-2.0 224 K from 1.6 GB
782 opus-mt-en-de Helsinki-NLP · translation Open weights — 512 ✓ cc-by-4.0 224 K —
783 gemma-2-9b-it google · text-generation · gated Open weights 9.2 B — ⚠ gemma 224 K from 22.2 GB
784 Ling-3.0-flash-GGUF AtomicChat · text-generation Open weights 127.5 B 262 K ✓ mit 224 K from 55.2 GB
785 resnet18.a3_in1k timm · image-classification Open weights 10 M — ✓ apache-2.0 224 K from 0.6 GB
786 Qwen3-14B-Base Qwen · text-generation Open weights 14.8 B 33 K ✓ apache-2.0 224 K from 35.2 GB
787 Qwen-Image-2.1-Lightning lightx2v · text-to-image Open weights 7.1 B — unknown 223 K from 1.6 GB
788 Qwen2-7B Qwen · text-generation Open weights 7.6 B 131 K ✓ apache-2.0 223 K from 18.4 GB
789 Agents-A1-4B-Q8_0-GGUF InternScience · text-generation Open weights — — ✓ apache-2.0 223 K from 1.2 GB
790 twitter-roberta-base-sentiment cardiffnlp · text-classification Open weights — 512 unknown 222 K —
791 t5-3b google-t5 · translation Open weights 2.9 B — ✓ apache-2.0 222 K from 13.5 GB
792 Ornith-1.0-35B-FP8 ornith-ai · text-generation Open weights 35.1 B — ✓ mit 222 K from 47.2 GB
793 MiniMax-M2.5-NVFP4 nvidia · text-generation Open weights 116.4 B 197 K other 222 K from 171.8 GB
794 JEV-35B-NVFP4 autotrust · text-classification Open weights 35.0 B — ✓ apache-2.0 222 K from 35.1 GB
795 Qwen2.5-72B-Instruct Qwen · text-generation Open weights Also via API 72.7 B 33 K other 221 K from 171.4 GB
796 MiniCPM5-1B-Claude-Opus-Fable5-V2-Thinking-GGUF GnLOLot · text-generation Open weights — — ✓ apache-2.0 221 K from 1.8 GB
797 LLaDA-8B-Instruct GSAI-ML · text-generation Open weights 8.0 B — ✓ mit 221 K from 19.3 GB
798 GLM-4.7-Flash-AWQ-4bit cyankiwi · text-generation Open weights 32.1 B 203 K ✓ mit 221 K from 27.5 GB
799 Olmo-3-7B-Instruct-SFT allenai · text-generation Open weights 7.3 B 66 K ✓ apache-2.0 220 K from 17.7 GB
800 MegaDescriptor-B-224 BVRA · image-classification Open weights — — ✗ cc-by-nc-4.0 220 K —
VRAM figures are estimates for the smallest available quantization at 8K context — see /methodology. Downloads refresh nightly from the Hugging Face API; API prices (blended 3:1 input:output per 1M tokens) and Artificial Analysis indices from OpenRouter.