§01·model · /models
gpt-oss 20B
llmactiveApache-2.0
20B open-weight LLM by OpenAI (gpt-oss) for local reasoning and agentic use. Apache-2.0.
§02·GPUs that run this model
23 total| GPU | VRAM | Series | Best speed | Min VRAM | Works | Benchmarks | Recipe | |
|---|---|---|---|---|---|---|---|---|
| RTX 4080 Super | 16GB | 40 | 6364prefill tokens/s | 16GB | ✓ | 1 | recipe | check ↗ |
| RTX 5090 | 32GB | 50 | 298.2tokens/s | ✓ | 3 | recipe | check ↗ | |
| RTX 5080 | 16GB | 50 | 172.4tokens/s | 16GB | ✓ | 2 | recipe | check ↗ |
| RTX 3090 Ti | 24GB | 30 | 160.3tokens/s | 24GB | ✓ | 2 | recipe | check ↗ |
| RTX 5070 Ti | 16GB | 50 | 156tokens/s | 16GB | ✓ | 2 | recipe | check ↗ |
| RTX 3090 | 24GB | 30 | 147.5tokens/s | 24GB | ✓ | 2 | recipe | check ↗ |
| RTX 4080 | 16GB | 40 | 136.5tokens/s | 16GB | ✓ | 2 | recipe | check ↗ |
| RTX 4070 Ti Super | 16GB | 40 | 128.9tokens/s | 16GB | ✓ | 2 | recipe | check ↗ |
| RTX 5060 Ti | 16GB | 50 | 92.1tokens/s | 16GB | ✓ | 2 | recipe | check ↗ |
| RTX 3060 | 12GB | 30 | 64tokens/s | ✓ | 1 | recipe | check ↗ | |
| RTX 4060 Ti 16GB | 16GB | 40 | 63.2tokens/s | 16GB | ✓ | 2 | recipe | check ↗ |
| Apple M2 Max | 64GB | apple | ~ | 0 | recipe | check ↗ | ||
| Apple M2 Pro | 16GB | apple | ~ | 0 | recipe | check ↗ | ||
| Apple M3 Max | 48GB | apple | ~ | 0 | recipe | check ↗ | ||
| Apple M4 Max | 48GB | apple | ~ | 0 | recipe | check ↗ | ||
| RTX 3080 Ti | 12GB | 30 | ~ | 0 | recipe | check ↗ | ||
| RTX 4070 | 12GB | 40 | ~ | 0 | recipe | check ↗ | ||
| RTX 4070 Super | 12GB | 40 | ~ | 0 | recipe | check ↗ | ||
| RTX 4070 Ti | 12GB | 40 | ~ | 0 | recipe | check ↗ | ||
| RTX 4090 | 24GB | 40 | ~ | 0 | recipe | check ↗ | ||
| RTX 5070 | 12GB | 50 | ~ | 0 | recipe | check ↗ | ||
| RX 7800 XT | 16GB | amd | ~ | 0 | recipe | check ↗ | ||
| RX 7900 XTX | 24GB | amd | ~ | 0 | recipe | check ↗ |
✓ benchmarked·~ runs via recipe (not benchmarked)·— untested·✕doesn't fit