RTX 5070 meets the minimum VRAM requirement for Q4 inference of openai/gpt-oss-20b. Review the quantization breakdown below to see how higher precision settings impact VRAM and throughput.
RTX 5070 can run openai/gpt-oss-20b with Q4 quantization. At approximately 60 tokens/second, you can expect Good speed - acceptable for interactive use.
VRAM usage will be very close to your GPU's limit. Consider closing other applications or using Q3 quantization for more margin.
| Quantization | VRAM needed | VRAM available | Estimated speed | Verdict |
|---|---|---|---|---|
| Q4 | 10GB | 12GB | 60.23 tok/s | ✅ Fits comfortably |
| Q8 | 20GB | 12GB | 42.10 tok/s | ❌ Not recommended |
| FP16 | 41GB | 12GB | 21.14 tok/s | ❌ Not recommended |