gemma-3-4b-it
by google
4.3B params · image-text-to-text · 1.4k likes · 1.6M downloads
gemma-3-4b-it is a 4.3B parameter model. At Q4 quantization it requires 2GB of VRAM. It runs comfortably on GeForce RTX 4090 (305 tok/s), GeForce RTX 5090 (457 tok/s), M4 Max 128GB (111 tok/s).
Inference providers
| Provider | $/1M in | $/1M out | Throughput |
|---|---|---|---|
| Featherless | |||
| deepinfra | 39 tok/s |
GPU compatibility
| GPU | VRAM | Q4 Decode | Verdict |
|---|---|---|---|
| GeForce RTX 4090 | 24GB | 305 tok/s | comfortable |
| GeForce RTX 5090 | 32GB | 457 tok/s | comfortable |
| M4 Max 128GB | 128GB | 111 tok/s | comfortable |
| M4 Pro 48GB | 48GB | 55 tok/s | comfortable |
| M4 Pro 24GB | 24GB | 55 tok/s | comfortable |
| A100 PCIe 80 GB | 80GB | 559 tok/s | comfortable |
| H100 SXM5 80 GB | 80GB | 1125 tok/s | comfortable |
| GeForce RTX 3090 | 24GB | 269 tok/s | comfortable |
| Radeon RX 7900 XTX | 24GB | 223 tok/s | comfortable |
| GeForce RTX 4080 | 16GB | 216 tok/s | comfortable |