AImodelsComparison.com 1,226 open · 200 API
Menu
Models Best of Trending Can I run it? Methodology About
1,226 models · 1,226 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 1,426 Open weights (self-host) 1,226 API-only 200
Task
All text-generation 534 text-to-image 155 image-text-to-text 112 ASR 86 sentence-similarity 64 translation 56 text-classification 51 feature-extraction 48 image-classification 39 token-classification 21 text-to-speech 19 audio-classification 12 image-segmentation 11 summarization 8 object-detection 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
$
1,226 models
#ModelParamsContextCommercial use30dMin VRAM
201 wav2vec2-base-vi-vlsp2020 nguyenvulebinh · ASR Open weights — — ✗ cc-by-nc-4.0 1.4 M —
202 chandra-ocr-2 datalab-to · image-text-to-text Open weights 5.3 B — ⚠ openrail 1.4 M from 12.9 GB
203 Swift-1.5-Qwen3.8-27B-GSQ-RCO-GGUF ukisai · text-generation Open weights 27.0 B — other 1.4 M from 4.6 GB
204 Qwen2.5-14B-Instruct-AWQ Qwen · text-generation Open weights 14.8 B 33 K ✓ apache-2.0 1.4 M from 13.7 GB
205 Qwen3-VL-8B-Instruct-FP8 Qwen · image-text-to-text Open weights 8.8 B — ✓ apache-2.0 1.4 M from 13.5 GB
206 DeepSeek-V3 deepseek-ai · text-generation Open weights Also via API 684.5 B 164 K unknown 1.3 M from 860.6 GB
207 Qwen3.5-9B-GGUF unsloth · image-text-to-text Open weights — — ✓ apache-2.0 1.3 M from 5.2 GB
208 NVIDIA-Nemotron-3-Nano-30B-A3B-NVFP4 nvidia · text-generation Open weights 18.2 B 262 K other 1.3 M from 24.5 GB
209 gender-classification rizvandwiki · image-classification Open weights 90 M — unknown 1.3 M from 0.9 GB
210 Qwen3-4B-Instruct-2507-FP8 Qwen · text-generation Open weights 4.4 B 262 K ✓ apache-2.0 1.3 M from 6.9 GB
211 NVIDIA-Nemotron-3-Nano-30B-A3B-BF16 nvidia · text-generation Open weights Also via API 31.6 B 262 K other 1.3 M from 74.7 GB
212 Dirk-Qwen3.8-27B-GGUF peculiar-ragdoll · image-text-to-text Open weights 27.0 B — ✓ apache-2.0 1.3 M from 14.2 GB
213 DeepSeek-R1-Distill-Qwen-1.5B deepseek-ai · text-generation Open weights 1.8 B 131 K ✓ mit 1.3 M from 4.7 GB
214 paraphrase-mpnet-base-v2 sentence-transformers · sentence-similarity Open weights 110 M 512 ✓ apache-2.0 1.3 M from 1.0 GB
215 ast-finetuned-audioset-10-10-0.4593 MIT · audio-classification Open weights 90 M — ✓ bsd-3-clause 1.3 M from 0.9 GB
216 Qwen3-VL-Embedding-8B Qwen · sentence-similarity Open weights 8.1 B — ✓ apache-2.0 1.3 M from 19.6 GB
217 table-transformer-structure-recognition microsoft · object-detection Open weights 30 M 1 K ✓ mit 1.3 M from 0.6 GB
218 wav2vec2-xls-r-300m-hebrew imvladikon · ASR Open weights 320 M — unknown 1.3 M from 1.9 GB
219 wav2vec2-large-xlsr-53-persian jonatasgrosman · ASR Open weights — — ✓ apache-2.0 1.3 M —
220 Qwen2.5-32B-Instruct Qwen · text-generation Open weights 32.8 B 33 K ✓ apache-2.0 1.3 M from 77.5 GB
221 Qwen3.6-27B-FP8 Qwen · image-text-to-text Open weights 27.8 B — ✓ apache-2.0 1.3 M from 38.6 GB
222 MiniCPM5-2B openbmb · text-generation Open weights 2.5 B 131 K ✓ apache-2.0 1.3 M from 6.4 GB
223 Qwen3.5-35B-A3B-FP8 Qwen · image-text-to-text Open weights 36.0 B — ✓ apache-2.0 1.2 M from 47.1 GB
224 Qwen3-ASR-1.7B Qwen · ASR Open weights 2.4 B — ✓ apache-2.0 1.2 M from 6.0 GB
225 wav2vec2-xls-r-parlaspeech-hr classla · ASR Open weights 320 M — unknown 1.2 M from 1.9 GB
226 turn-detector livekit · text-classification Open weights 130 M 8 K other 1.2 M from 1.1 GB
227 Qwen3.6-35B-A3B-GGUF unsloth · image-text-to-text Open weights — — ✓ apache-2.0 1.2 M from 11.6 GB
228 Qwen3.5-4B-GGUF unsloth · image-text-to-text Open weights — — ✓ apache-2.0 1.2 M from 2.8 GB
229 Qwen3.8-27B-GSQ-RCO-GGUF ISTA-DASLab · image-text-to-text Open weights — — ✓ apache-2.0 1.2 M from 1.5 GB
230 faster-whisper-large-v3 Systran · ASR Open weights — — ✓ mit 1.2 M —
231 Qwen2.5-0.5B Qwen · text-generation Open weights 490 M 33 K ✓ apache-2.0 1.2 M from 1.7 GB
232 nb-wav2vec2-1b-bokmaal-v2 NbAiLab · ASR Open weights 960 M — ✓ apache-2.0 1.2 M from 4.9 GB
233 bart-large-cnn facebook · summarization Open weights 410 M 1 K ✓ mit 1.2 M from 2.3 GB
234 faster-whisper-medium Systran · ASR Open weights — — ✓ mit 1.2 M —
235 Qwen3.8-35B-A3B-Distill-GGUF empero-ai · text-generation Open weights 35.0 B — ✓ apache-2.0 1.2 M from 19.6 GB
236 Ornith-1.5-397B-GGUF ornith-ai · text-generation Open weights — — ✓ mit 1.2 M from 1.5 GB
237 Wav2Vec2-large-xlsr-hindi theainerd · ASR Open weights 316 M — unknown 1.2 M from 1.9 GB
238 GEV-26B-Decide autotrust · text-classification Open weights 25.8 B — ✓ apache-2.0 1.2 M from 61.4 GB
239 opus-mt-nl-en Helsinki-NLP · translation Open weights — 512 ✓ apache-2.0 1.2 M —
240 kokoro-inno-clone-tuner remsky · text-to-speech Open weights 9.7 M — ✓ apache-2.0 1.2 M from 0.5 GB
241 rtdetr_r101vd_coco_o365 PekingU · object-detection Open weights 80 M — ✓ apache-2.0 1.2 M from 0.9 GB
242 bert-base-NER dslim · token-classification Open weights 110 M 512 ✓ mit 1.1 M from 1.0 GB
243 faster-whisper-small.en Systran · ASR Open weights — — ✓ mit 1.1 M —
244 nemotron-3.5-asr-streaming-0.6b nvidia · ASR Open weights 640 M — other 1.1 M from 1.4 GB
245 stable-diffusion-v1-5 stable-diffusion-v1-5 · text-to-image Open weights 860 M — ⚠ creativeml-openrail-m 1.1 M from 26.6 GB
246 cohere-transcribe-03-2026-gguf handy-computer · ASR Open weights 2.1 B — ✓ apache-2.0 1.1 M from 2.5 GB
247 jina-embeddings-v2-small-en jinaai · feature-extraction Open weights 30 M 8 K ✓ apache-2.0 1.1 M from 0.6 GB
248 gemma-4-26B-A4B-it-FP8-dynamic RedHatAI · image-text-to-text Open weights 26.5 B — ✓ apache-2.0 1.1 M from 36.0 GB
249 Qwen3-Coder-30B-A3B-Instruct-FP8 Qwen · text-generation Open weights 30.5 B 262 K ✓ apache-2.0 1.1 M from 39.4 GB
250 Ornith-1.5-35B-A3B-NVFP4 ornith-ai · text-generation Open weights 19.5 B — ✓ mit 1.1 M from 29.2 GB
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.