AImodelsComparison.com 1,213 open · 210 API
Menu
Models Best of Trending Can I run it? Methodology MCP server About
1,423 models · 1,213 open weights · 210 API-only · refreshed nightly

Models that run on RTX 4090 · 24 GB

Every model in the catalog — open weights you can self-host and API-only models — with licence or API price, estimated VRAM and daily-tracked downloads. Filters update the URL — share any view.

Type
All 1,423 Open weights (self-host) 1,213 API-only 210
Task
All text-generation 735 text-to-image 165 image-text-to-text 110 ASR 82 sentence-similarity 63 text-classification 50 translation 50 feature-extraction 46 image-classification 41 text-to-speech 23 token-classification 22 summarization 9 audio-classification 9 image-segmentation 9 object-detection 7 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
$
762 models
#ModelParamsContextCommercial use30dOn RTX 4090 · 24 GB
601 VLM2Vec-Full TIGER-Lab · text-generation Open weights 4.2 B 131 K ✓ apache-2.0 199 K 10.2 GB
602 Qwen2.5-7B-Instruct-1M Qwen · text-generation Open weights 7.6 B 1.0 M ✓ apache-2.0 197 K 18.4 GB
603 Olmo-3-7B-Think allenai · text-generation Open weights 7.3 B 66 K ✓ apache-2.0 197 K 17.7 GB
604 gemma-4-12b-heretic-abliterated-GGUF culturerevolt · text-generation Open weights — — ✓ apache-2.0 196 K 0.7 GB
605 Olmo-3-7B-Instruct allenai · text-generation Open weights 7.3 B 66 K ✓ apache-2.0 196 K 17.7 GB
606 Qwen2.5-Coder-32B-Instruct-GGUF Qwen · text-generation Open weights — — ✓ apache-2.0 196 K 9.1 GB
607 llama-3-8b-instruct-awq casperhansen · text-generation Open weights 8.0 B 8 K unknown 196 K 8.0 GB
608 Fanar-1-9B-Instruct QCRI · text-generation Open weights 8.8 B 4 K ✓ apache-2.0 195 K 21.1 GB
609 Qwen2.5-3B-Instruct-unsloth-bnb-4bit unsloth · text-generation Open weights 3.2 B 33 K ✓ apache-2.0 195 K 3.6 GB
610 Qwen3-1.7B-GGUF unsloth · text-generation Open weights 1.7 B 41 K ✓ apache-2.0 195 K 1.6 GB
611 gpt-oss-120b-Eagle3-short-context nvidia · text-generation Open weights 787 M 131 K other 195 K 2.8 GB
612 Qwen2.5-1.5B-Instruct-AWQ Qwen · text-generation Open weights 1.8 B 33 K ✓ apache-2.0 195 K 2.5 GB
613 Llama-Guard-3-8B meta-llama · text-generation · gated Open weights 8.0 B — ⚠ llama3.1 195 K 19.4 GB
614 NVIDIA-Nemotron-3.5-Lightning-30B-A3B-NVFP4-DSpark nvidia · text-generation Open weights 760 M 1.0 M other 195 K 2.1 GB
615 TinyLlama-1.1B-Chat-v0.3-GPTQ TheBloke · text-generation Open weights 1.1 B 2 K ✓ apache-2.0 195 K 1.5 GB
616 Illustrious-xl-early-release-v0 OnomaAIResearch · text-to-image Open weights 2.6 B — other 194 K 23.8 GB
617 Qwen3.6-35B-A3B-Claude-4.6-Opus-Reasoning-Distilled-GGUF hesamation · text-generation Open weights — — ✓ apache-2.0 194 K 23.8 GB
618 EuroLLM-1.7B-Instruct utter-project · text-generation Open weights 1.7 B 4 K ✓ apache-2.0 194 K 4.4 GB
619 supergemma4-26b-uncensored-gguf-v2 Jiunsong · text-generation Open weights — — ⚠ gemma 194 K 19.0 GB
620 gemma-2-2b google · text-generation · gated Open weights 2.6 B — ⚠ gemma 193 K 12.4 GB
621 Qwen2.5-Coder-14B-Instruct-GGUF Qwen · text-generation Open weights 14.0 B 33 K ✓ apache-2.0 192 K 8.9 GB
622 Llama-3.2-3B-Instruct-GGUF bartowski · text-generation Open weights — — ⚠ llama3.2 191 K 2.3 GB
623 Qwen3.8-27B-DSpark RadixArk · text-generation Open weights 1.9 B 262 K other 190 K 4.9 GB
624 NVIDIA-Nemotron-Nano-9B-v2-FP8 nvidia · text-generation Open weights 8.9 B 131 K other 190 K 13.1 GB
625 Qwen3-14B-unsloth-bnb-4bit unsloth · text-generation Open weights 15.2 B 41 K ✓ apache-2.0 190 K 15.0 GB
626 DeepSeek-R1-0528-Qwen3-8B-MLX-8bit lmstudio-community · text-generation Open weights 2.3 B 131 K ✓ mit 189 K 10.4 GB
627 nova-furry-xl-il-v120-sdxl John6666 · text-to-image Open weights 2.6 B — other 189 K 8.5 GB
628 gemma-3-1b-it-GGUF lmstudio-community · text-generation Open weights 1.0 B — ⚠ gemma 187 K 1.5 GB
629 LFM2.5-2.6B-GGUF LiquidAI · text-generation Open weights 2.6 B 131 K other 187 K 2.6 GB
630 Qwen3.6-14B-A3B-FableVibes-GGUF tvall43 · text-generation Open weights — — ✓ apache-2.0 187 K 6.4 GB
631 opt-350m facebook · text-generation Open weights 350 M 2 K other 185 K 1.3 GB
632 Qwen_Qwen3-0.6B-GGUF bartowski · text-generation Open weights 600 M 41 K unknown 185 K 1.0 GB
633 Qwen-Image-2.1-viggle-turbo Viggle · text-to-image Open weights 7.1 B — other 183 K 6.3 GB
634 FLUX.2-klein-4B-GGUF leejet · text-to-image Open weights — — ✓ apache-2.0 182 K 3.2 GB
635 phi-2 microsoft · text-generation Open weights 2.8 B 2 K ✓ mit 178 K 7.0 GB
636 Qwen3-1.7B-GGUF Qwen · text-generation Open weights 1.7 B 41 K ✓ apache-2.0 176 K 2.8 GB
637 Qwen3.8-4B-Distill empero-ai · text-generation Open weights 4.7 B — ✓ apache-2.0 175 K 11.5 GB
638 Ternary-Bonsai-8B-gguf prism-ml · text-generation Open weights — — ✓ apache-2.0 174 K 2.9 GB
639 Underdog-Saluki-27B-1.0-MTP-GGUF Kujira · text-generation Open weights 27.0 B — ✓ apache-2.0 173 K 13.7 GB
640 t5gemma-2b-2b-ul2-it google · text-generation · gated Open weights 5.6 B — ⚠ gemma 173 K 13.7 GB
641 deepseek-coder-7b-instruct-v1.5 deepseek-ai · text-generation Open weights 6.9 B 4 K other 172 K 16.7 GB
642 gpt-oss-20b-MXFP4-Q8 mlx-community · text-generation Open weights 20.9 B 131 K ✓ apache-2.0 171 K 16.9 GB
643 Meta-Llama-3.1-8B-Instruct unsloth · text-generation Open weights 8.0 B 131 K ⚠ llama3.1 169 K 19.4 GB
644 Qwen-AgentWorld-35B-A3B-GGUF unsloth · text-generation Open weights — — ✓ apache-2.0 169 K 13.1 GB
645 Realistic_Vision_V6.0_B1_noVAE SG161222 · text-to-image Open weights — — ⚠ creativeml-openrail-m 168 K 14.6 GB
646 MiniCPM5-1B-Claude-Opus-Fable5-V2-Thinking-GGUF GnLOLot · text-generation Open weights — — ✓ apache-2.0 168 K 1.8 GB
647 gpt2-medium openai-community · text-generation Open weights 380 M — ✓ mit 166 K 2.2 GB
648 Qwen3-0.6B-GGUF unsloth · text-generation Open weights 600 M 41 K ✓ apache-2.0 165 K 0.9 GB
649 FLUX.1-schnell-gguf city96 · text-to-image Open weights — — ✓ apache-2.0 165 K 4.9 GB
650 Qwen2-0.5B Qwen · text-generation Open weights 490 M 131 K ✓ apache-2.0 165 K 1.7 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.