What does your model need?
Find an open model, see the GPU memory it takes to run or fine-tune, and the cheapest GPU that fits.
652 models. Memory marked ~ is an estimate from the model's size. Prices are full price for one GPU in stock, without VAT.
| Model | Task | GPU memory | Cheapest GPU that fits | Parameters | Downloads | License |
|---|---|---|---|---|---|---|
| Kimi K3.1 Pro moonshotai | Image text to text | 1874 GB | — | 2780B | 27.9K | Other |
| Qwen3.8 Max (2.4T) Qwen | Text generation | 5871 GB | — | 2446B | 26.3K | Other |
| Wan 2.2 T2V A14B Wan-AI | video | 80 GB | €2.09/h RTX PRO 6000 | 27B | — | Apache 2.0 |
| MiniMax M3 MiniMaxAI | llm | 1025 GB | — | 428B | — | MiniMax Community License |
| FLUX.2 [dev] black-forest-labs | image | 64 GB | €2.09/h RTX PRO 6000 | 32.2B | — | FLUX [dev] Non-Commercial License |
| MiniMax H3 MiniMaxAI | video | 202 GB | — | 33B | — | MiniMax H3 Community License |
| Wan 2.2 TI2V 5B Wan-AI | video | 24 GB | €0.17/h RTX 3090 | 5B | — | Apache 2.0 |
| all-MiniLM-L6-v2 sentence-transformers | embedding | 0.1 GB | €0.04/h RTX 3060 | 0B | — | Apache 2.0 |
| LTX-2.5 Lightricks | video | 32 GB | €0.62/h RTX 6000 Ada | 19B | — | LTX-2.x Community License |
| Kimi K2.7 Code moonshotai | coding | 714 GB | — | 1000B | — | Modified MIT |
| HunyuanVideo 1.5 tencent | video | 14 GB | €0.17/h RTX 3090 | — | — | Tencent Hunyuan Community License |
| Mistral Large 4 mistralai | llm | 1273 GB | — | 1000B | — | Mistral Ecosystem Open-Weight License (modified MIT) |
| FLUX.2 [klein] 4B black-forest-labs | image | 13 GB | €0.17/h RTX 3090 | 3.9B | — | Apache 2.0 |
| Qwen3-Coder 480B Qwen | coding | 1152 GB | — | 480B | — | Apache 2.0 |
| Devstral 2 123B mistralai | coding | 143 GB | — | 125B | — | Modified MIT |
| Qwen3-Coder-Next Qwen | coding | 178 GB | — | 79.7B | — | Apache 2.0 |
| BGE Reranker v2 M3 BAAI | embedding | 2.5 GB | €0.04/h RTX 3060 | 0.6B | — | Apache 2.0 |
| BGE-M3 BAAI | embedding | 2.5 GB | €0.04/h RTX 3060 | — | — | MIT |
| Kokoro 82M hexgrad | speech | 0.4 GB | €0.04/h RTX 3060 | 0.1B | — | Apache 2.0 |
| FLUX 3 Image [klein] black-forest-labs | image | 24 GB | €0.17/h RTX 3090 | 5.5B | — | FLUX Kommunity License |
| DeepSeek V4 Pro (0813) deepseek-ai | llm | 960 GB | — | 1600B | — | MIT |
| bge-small-en-v1.5 BAAI | Feature extraction | ~0.1 GB | €0.04/h RTX 3060 | 0B | 47.6M | MIT |
| gpt2 openai-community | Text generation | ~0.3 GB | €0.04/h RTX 3060 | 0.1B | 31.8M | MIT |
| Qwen3-0.6B Qwen | Text generation | ~1.7 GB | €0.04/h RTX 3060 | 0.8B | 20.7M | Apache 2.0 |
| Qwen3-14B Qwen | Text generation | ~33 GB | €0.62/h RTX 6000 Ada | 14.8B | 12.9M | Apache 2.0 |