models · nvidia
Nemotron 3.5 Lightning — catalogued from third-party evaluation data; no weights held by Sovereign Frontier.
| Method | hf-published-sha256 |
|---|
| Publisher | nvidia |
|---|---|
| Origin | United States |
| Licence | NVIDIA-Open-Model |
| Publisher’s own licence statement | other |
| Parameters | 31,577,937,344 (31.6B) |
| Context window | 1,000,000 tokens |
| Modality | text->text |
| Architecture | nemotron_h |
| Layers | 52 |
| Hidden size | 2688 |
| Attention | GQA (32 heads, 2 KV) |
| Native precision | BF16 |
| Downloads reported upstream | 22,279 |
| GPQA Diamond | 74.3 |
|---|---|
| Humanity's Last Exam | 10.6 |
Reported by a third party and recorded here, not re-run by us. Source: artificialanalysis.ai/api/v2.
| Hardware | Engine | Build measured | Decode, 1 stream | Label | Reported by |
|---|---|---|---|---|---|
| Intel Arc Pro B70 | llama.cpp | Q4_0 | 87.4 tok/s @ 16,384 | candidate | localmaxxing |
| Intel Arc Pro B70 | llama.cpp | Q4_K_M | 49.8 tok/s @ 16,384 | candidate | localmaxxing |
| GeForce RTX 4070 | llama.cpp | GGUF Q4_K_M | 43.5 tok/s @ 4,096 | candidate | localmaxxing |
From local-ai-registry (MIT), commit 124959d: 11 runs across 5 machines, 0 validated — model revision and runtime pinned, launch accepted — and 11 candidate, their label for useful evidence without a reproducible-launch promise.
Registry record as of 2026-08-13. 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.