madlad400-3b-mt
by google
2.9B params · translation · 208 likes · 197.4k downloads
madlad400-3b-mt is a 2.9B parameter model. At Q4 quantization it requires 1GB of VRAM. It runs comfortably on GeForce RTX 4090 (446 tok/s), GeForce RTX 5090 (669 tok/s), M4 Max 128GB (163 tok/s).
Inference providers
| Provider | $/1M in | $/1M out | Throughput |
|---|---|---|---|
| HF Inference |
GPU compatibility
| GPU | VRAM | Q4 Decode | Verdict |
|---|---|---|---|
| GeForce RTX 4090 | 24GB | 446 tok/s | comfortable |
| GeForce RTX 5090 | 32GB | 669 tok/s | comfortable |
| M4 Max 128GB | 128GB | 163 tok/s | comfortable |
| M4 Pro 48GB | 48GB | 81 tok/s | comfortable |
| M4 Pro 24GB | 24GB | 81 tok/s | comfortable |
| A100 PCIe 80 GB | 80GB | 818 tok/s | comfortable |
| H100 SXM5 80 GB | 80GB | 1645 tok/s | comfortable |
| GeForce RTX 3090 | 24GB | 394 tok/s | comfortable |
| Radeon RX 7900 XTX | 24GB | 326 tok/s | comfortable |
| GeForce RTX 4080 | 16GB | 316 tok/s | comfortable |