AImodelsComparison.com 1,000 models
Menu
Models Best of Trending Can I run it? Methodology About
1,000 models · refreshed nightly

Models that run on RTX 4090 · 24 GB

Every model in the catalog with its licence, estimated VRAM and daily-tracked downloads. Filters update the URL — share any view.

Task
All text-generation 400 text-to-image 150 ASR 85 image-text-to-text 80 sentence-similarity 54 feature-extraction 46 translation 43 text-classification 39 image-classification 32 token-classification 21 text-to-speech 15 audio-classification 12 image-segmentation 8 object-detection 8 summarization 7
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
$
636 models
#ModelParamsContextCommercial use30dOn RTX 4090 · 24 GB
401 granite-4.1-3b ibm-granite · text-generation 3.4 B 131 K ✓ apache-2.0 312 K 8.5 GB
402 NVIDIA-Nemotron-Nano-9B-v2 nvidia · text-generation 8.9 B 131 K other 310 K 21.4 GB
403 RMBG-1.4 briaai · image-segmentation 40 M — other 310 K 0.7 GB
404 segformer_b2_clothes mattmdjaga · image-segmentation 30 M — other 308 K 0.6 GB
405 voice-gender-classifier JaesungHuh · audio-classification 20 M — ✓ mit 307 K 0.6 GB
406 Jan-v3.5-4B-gguf janhq · text-generation — — ✓ apache-2.0 306 K 2.8 GB
407 Qwen2.5-3B Qwen · text-generation 3.1 B 33 K other 306 K 7.8 GB
408 xlm-roberta-base-ner-hrl Davlan · token-classification 280 M 512 ✓ afl-3.0 304 K 1.8 GB
409 Olmo-3-7B-Instruct-SFT allenai · text-generation 7.3 B 66 K ✓ apache-2.0 301 K 17.7 GB
410 phishing-email-detection-distilbert_v2.4.1 cybersectony · text-classification 70 M 512 ✓ apache-2.0 301 K 0.8 GB
411 Qwen2-0.5B-Instruct Qwen · text-generation 490 M 33 K ✓ apache-2.0 294 K 1.7 GB
412 Meta-Llama-3.1-8B-Instruct-GGUF bartowski · text-generation — — ⚠ llama3.1 294 K 3.7 GB
413 mms-lid-126 facebook · audio-classification 970 M — ✗ cc-by-nc-4.0 291 K 4.9 GB
414 deberta-v3-base-prompt-injection-v2 protectai · text-classification 180 M 512 ✓ apache-2.0 291 K 1.3 GB
415 CommunityForensics-DeepfakeDet-ViT buildborderless · image-classification 40 M — ✓ mit 290 K 0.7 GB
416 Meta-Llama-3-8B NousResearch · text-generation 8.0 B 8 K other 290 K 19.4 GB
417 rtdetr_v2_r18vd PekingU · object-detection 20 M — ✓ apache-2.0 289 K 0.6 GB
418 SmolLM-135M HuggingFaceTB · text-generation 130 M 2 K ✓ apache-2.0 288 K 1.1 GB
419 DeepSeek-R1-Distill-Qwen-7B deepseek-ai · text-generation 7.6 B 131 K ✓ mit 288 K 18.4 GB
420 Llama-3.2-1B-Instruct unsloth · text-generation 1.2 B 131 K ⚠ llama3.2 287 K 3.4 GB
421 MuQ-large-msd-iter OpenMuQ · audio-classification 330 M — ✗ cc-by-nc-4.0 286 K 2.0 GB
422 gemma-4-12B-coder-fable5-composer2.5-v1-GGUF yuxinlu1 · text-generation — — ✓ apache-2.0 286 K 5.8 GB
423 roberta-large-mnli FacebookAI · text-classification 360 M 512 ✓ mit 285 K 2.1 GB
424 llama-7b huggyllama · text-generation 6.7 B 2 K other 285 K 16.3 GB
425 MiMo-7B-Base XiaomiMiMo · text-generation 7.8 B 33 K ✓ mit 284 K 18.9 GB
426 vit_small_patch16_224.augreg_in21k_ft_in1k timm · image-classification 20 M — ✓ apache-2.0 283 K 0.6 GB
427 Qwen2.5-7B-Instruct-GPTQ-Int4 Qwen · text-generation 7.6 B 33 K ✓ apache-2.0 282 K 7.8 GB
428 nsfw-classifier giacomoarienti · image-classification 90 M — ✗ cc-by-nc-nd-4.0 279 K 0.9 GB
429 granite-4.0-h-tiny ibm-granite · text-generation 6.9 B 131 K ✓ apache-2.0 279 K 16.8 GB
430 DeepSeek-R1-0528-Qwen3-8B-MLX-4bit lmstudio-community · text-generation 1.3 B 131 K ✓ mit 278 K 5.8 GB
431 Qwen3-8B-GGUF unsloth · text-generation — 41 K ✓ apache-2.0 277 K 4.1 GB
432 open-vakgyata onecxi · audio-classification 60 M — ✗ cc-by-nc-4.0 274 K 0.8 GB
433 Qwen3-4B-GGUF Qwen · text-generation — — ✓ apache-2.0 274 K 3.2 GB
434 MiniCPM5-1B-Claude-Opus-Fable5-V2-Thinking-GGUF GnLOLot · text-generation — — ✓ apache-2.0 265 K 1.8 GB
435 Qwen2.5-Math-1.5B Qwen · text-generation 1.5 B 4 K ✓ apache-2.0 265 K 4.1 GB
436 madlad400-3b-mt google · translation 2.9 B — ✓ apache-2.0 263 K 2.8 GB
437 gemma-2-2b google · text-generation · gated 2.6 B — ⚠ gemma 263 K 12.4 GB
438 wikineural-multilingual-ner Babelscape · token-classification 180 M 512 ✗ cc-by-nc-sa-4.0 263 K 1.3 GB
439 DeepSeek-R1-0528-Qwen3-8B-MLX-8bit lmstudio-community · text-generation 2.3 B 131 K ✓ mit 258 K 10.4 GB
440 plant-identity umutbozdag · image-classification 90 M — unknown 257 K 0.9 GB
441 Qwen3-Coder-30B-A3B-Instruct-AWQ QuantTrio · text-generation 30.5 B 262 K ✓ apache-2.0 257 K 23.6 GB
442 t5-3b google-t5 · translation 2.9 B — ✓ apache-2.0 255 K 13.5 GB
443 Sugoi-32B-Ultra-GGUF sugoitoolkit · translation — — ✓ apache-2.0 254 K 14.0 GB
444 OLMoE-1B-7B-0125-Instruct allenai · text-generation 6.9 B 4 K ✓ apache-2.0 251 K 16.8 GB
445 bert-base-multilingual-cased-ner-hrl Davlan · token-classification 180 M 512 ✓ afl-3.0 249 K 1.3 GB
446 span-marker-bert-base-uncased-acronyms tomaarsen · token-classification 110 M — ✓ apache-2.0 247 K 1.0 GB
447 Bielik-11B-v3.0-Instruct-awq speakleash · text-generation 11.3 B 33 K ✓ apache-2.0 247 K 9.0 GB
448 Mistral-Small-24B-Instruct-2501-AWQ stelterlab · text-generation 23.6 B 33 K ✓ apache-2.0 246 K 19.7 GB
449 xlm-roberta-large-ner-hrl Davlan · token-classification 560 M 512 ✓ afl-3.0 246 K 3.0 GB
450 wav2vec2-large-robust-24-ft-age-gender audeering · audio-classification 320 M — ✗ cc-by-nc-sa-4.0 246 K 1.9 GB
VRAM figures are estimates for the smallest available quantization at 8K context — see /methodology. Downloads refresh nightly from the Hugging Face API.