Loading…
Loading…
Models / Qwen3-Embedding
GPU memory needed (smallest version of each): Qwen3-Embedding 0.6B about 2.9 GB, Qwen3-Embedding 4B about 11.2 GB, Qwen3-Embedding 8B about 19.7 GB. The cheapest runs on RTX 3060 from $0.09/h right now.
3 models (0.6B to 8B) · ↓ 12.4K downloads a month · from $0.09/h on RTX 3060 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 |
|---|---|---|---|---|
| Originalunsloth/Qwen3-Embedding-0.6B | vLLM | 1.2 GB | 2.9 GB | $0.09/h RTX 3060 |
| Qwen3 Embedding Medical 0.6B (lastmass)lastmass/Qwen3-Embedding-Medical-0.6B | vLLM | 1.3 GB | 3 GB | $0.09/h RTX 3060 |
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|---|---|---|---|
| Originalunsloth/Qwen3-Embedding-4B | vLLM | 8.0 GB | 11.2 GB | $0.10/h RTX 3060 |
| Version | Runtime | Weights | GPU memory | Cheapest machine now |
|---|---|---|---|---|
| Originalunsloth/Qwen3-Embedding-8B | vLLM | 15 GB | 19.7 GB | $0.21/h RTX 4000 Ada |
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").