Loading…
Loading…
Models / Qwen3
4 models (0.6B to 1.7B) · ↓ 909.2K downloads a month · from $0.08/h on GTX 1660 SUPER right now. Each model's versions do the same job; quantized ones (AWQ, FP8, 4-bit) need less GPU memory and run on cheaper machines.
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|---|---|---|---|
| OriginalPrimeIntellect/Qwen3-0.6B | vLLM | 1.5 GB | 4.9 GB | $0.10/h RTX 3060 |
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|---|---|---|---|
| ONNX (haydso)haydso/Qwen3-0.6B-Reverse-Text-SFT-ONNX | ONNX Runtime GenAI | 710 MB | 3.3 GB | $0.08/h GTX 1660 SUPER |
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|---|---|---|---|
| OriginalPrimeIntellect/Qwen3-1.7B | vLLM | 4.1 GB | 7.8 GB | $0.10/h RTX 3060 |
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|---|---|---|---|
| ONNX (haydso)haydso/Qwen3-1.7B-Wordle-SFT-ONNX | ONNX Runtime GenAI | 1.7 GB | 4.4 GB | $0.08/h GTX 1660 SUPER |
Model details come from Hugging Face and are refreshed daily; prices follow the live offers. GPU memory is GridPup's estimate: the weights, room for an 8,192-token conversation (chat models) and the runtime's own memory. Large models are split across the GPUs of one machine ("4x RTX 4090").