feat(deepinfra): add new models [bot]

[harshiv-26]

Auto-generated by model-addition-agent for provider deepinfra.

---

Note

Low Risk
Low risk: this PR only adds new DeepInfra model definition YAMLs (pricing/limits/modalities) without changing runtime logic.

Overview
Adds two new DeepInfra model configuration entries: ByteDance/Seed-2.0-code (with prompt_caching and cache-read token pricing) and nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning.

Each definition specifies token costs, very large context/max token limits (~256k), and declares image input modality, expanding the set of selectable DeepInfra models.

^{Reviewed by Cursor Bugbot for commit 1849211. Bugbot is set up for automated code reviews on this repo. Configure here.}

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 1849211. Configure here.}

[cursor]

Missing text input modality for text-based models

High Severity

Both new model configs list only image as their input modality, missing text. A code model (Seed-2.0-code) and a reasoning model (Nemotron-3-Nano-Omni-30B-A3B-Reasoning) fundamentally require text input. All sibling models (e.g., Seed-2.0-pro, Seed-2.0-mini, Seed-1.8, Nemotron-3-Nano-30B-A3B) include text in their input modalities. This omission would cause these models to be incorrectly classified as image-only, potentially preventing text-based usage.

Additional Locations (1)

• providers/deepinfra/nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning.yaml#L7-L10

 

^{Reviewed by Cursor Bugbot for commit 1849211. Configure here.}