models · ornith-ai
Catalogued from local.ai local-inference benchmark index at Quarantine — unverified, pending attestation.
| Method | hf-published-sha256 |
|---|
| Publisher | ornith-ai |
|---|---|
| Origin | Not recorded. The harvest wrote its own source name into this field, which tells us nothing about jurisdiction. |
| Licence | MIT |
| Parameters | 35,107,181,936 (35.1B) |
| Native precision | BF16 |
| Downloads reported upstream | 3,286,288 |
| Hardware | Build measured | Decode @ 8K | Memory @ 8K | Decode mode |
|---|---|---|---|---|
| NVIDIA RTX PRO 6000 Blackwell 96GB | Q4_K_M | 271 tok/s | 21.9 GB | ordinary |
| NVIDIA GeForce RTX 5090 32GB | nvfp4 | 234 tok/s | 29.5 GB | ordinary |
| NVIDIA RTX PRO 6000 Blackwell 96GB | nvfp4 | 232 tok/s | 87.4 GB | ordinary |
| NVIDIA RTX PRO 6000 Blackwell 96GB | FP8 | 195 tok/s | 90.7 GB | ordinary |
| NVIDIA RTX 6000 Ada 48GB | Q4_K_M | 188 tok/s | 26.5 GB | ordinary |
| NVIDIA RTX 6000 Ada 48GB | nvfp4 | 178 tok/s | 42.9 GB | ordinary |
| NVIDIA RTX 6000 Ada 48GB | FP8 | 132 tok/s | 44.6 GB | ordinary |
| MacBook Pro M5 Max 128GB · 40-core GPU | Q4_K_M | 113 tok/s | 27.1 GB | ordinary |
Measured by local.ai, not by us: 16 runs, 8 more not listed. Each names its engine pinned by image digest and the exact serve command, in the full record.
Registry record as of 2026-09-26. Identity, licence and parameter count are properties of a release and do not move; tier can, and the verify link above re-checks it against the live registry rather than this page.