roberta-base-openai-detector
by openai-community
125M params · text-classification · 131 likes · 1.1M downloads
roberta-base-openai-detector is a 125M parameter model. At Q4 quantization it requires 0GB of VRAM. It runs comfortably on GeForce RTX 4090 (10484 tok/s), GeForce RTX 5090 (15722 tok/s), M4 Max 128GB (3836 tok/s).
Inference providers
| Provider | $/1M in | $/1M out | Throughput |
|---|---|---|---|
| HF Inference |
GPU compatibility
| GPU | VRAM | Q4 Decode | Verdict |
|---|---|---|---|
| GeForce RTX 4090 | 24GB | 10484 tok/s | comfortable |
| GeForce RTX 5090 | 32GB | 15722 tok/s | comfortable |
| M4 Max 128GB | 128GB | 3836 tok/s | comfortable |
| M4 Pro 48GB | 48GB | 1918 tok/s | comfortable |
| M4 Pro 24GB | 24GB | 1918 tok/s | comfortable |
| A100 PCIe 80 GB | 80GB | 19208 tok/s | comfortable |
| H100 SXM5 80 GB | 80GB | 38633 tok/s | comfortable |
| GeForce RTX 3090 | 24GB | 9269 tok/s | comfortable |
| Radeon RX 7900 XTX | 24GB | 7665 tok/s | comfortable |
| GeForce RTX 4080 | 16GB | 7440 tok/s | comfortable |