Add NVIDIA Nemotron-3-Ultra (subscription) parameters

[guillaumegay13]

What this changes

Adds parameters for NVIDIA Nemotron-3-Ultra on the subscription route (Ollama Cloud, $20/mo), filed as #74. The API-key entry already exists as nemotron-3-ultra-550b-a55b; this adds the subscription variant under the model id Ollama Cloud serves (nemotron-3-ultra).

Mirrors the existing nemotron-3-ultra-550b-a55b entry's user-facing toggles: temperature, top_p, max_tokens, reasoning_effort (none/medium/high, default high), reasoning_budget, and stop.

Type of change

• Add a model (new YAML file under models/)

Source

• Model card: https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16
• Param shape mirrors the existing models/nvidia/nemotron-3-ultra-550b-a55b.yaml (same model, API-key route).
• Available on Ollama Cloud as nemotron-3-ultra (confirmed via the live model listing).

Smoke-tested live through a local Manifest server. The Ollama Cloud key on hand is expired, so the sampling + reasoning params were exercised against the same model on OpenRouter (nvidia/nemotron-3-ultra-550b-a55b:free): all accepted (200).

Closes #74

[vercel]

@guillaumegay13 is attempting to deploy a commit to the Manifest Team on Vercel.

A member of the Team first needs to authorize it.

[guillaumegay13]

Superseded by #80, which bundles the remaining model-param requests into a single PR so the merge cuts one modelparams release instead of four. Same files, same live smoke tests.