update model intance to nano

[cdgamarose-nv]

No description provided.

[greptile-apps]

Greptile Summary

This PR updates the default LLM used in the Python API "simplest run" demo cell of 1_Deep_Researcher_Web_Search.ipynb from nvidia/nemotron-3-super-120b-a12b (120B parameter Super model) to nvidia/nemotron-3-nano-30b-a3b (30B parameter Nano model), aligning it with the YAML configs already present in the same notebook which use the nano model for the researcher role and comment out Super as "limited availability."

• The change is a single-line model name swap in one notebook cell; the YAML-based configs (config_simple_deep_researcher.yml and config_deep_research_bench.yml) embedded in the notebook already referenced the nano model and are unaffected.
• The Python cell uses the nano model for all roles (orchestrator, planner, researcher), while the YAML configs assign nano only to the researcher role and use gpt_oss_llm for orchestrator and planner — this is intentional per the "One LLM for all roles" comment and serves as a minimal getting-started example.
• The YAML configs include chat_template_kwargs: enable_thinking: true for the nano model but the Python cell does not pass an equivalent parameter, creating a minor inconsistency in how the model is invoked between the two execution paths.

Confidence Score: 4/5

• Safe to merge — isolated single-line model name update in a notebook demo cell with no production code impact.
• The change is minimal (one model name string) in a documentation notebook. The nano model is already established elsewhere in the same file. The only minor concern is a missing enable_thinking flag in the Python cell that is present in the YAML configs, which could produce slightly different model behaviour between the two demo paths but does not affect correctness or production systems.
• No files require special attention.

Important Files Changed

Filename	Overview
docs/notebooks/1_Deep_Researcher_Web_Search.ipynb	Single-line model swap from nvidia/nemotron-3-super-120b-a12b to nvidia/nemotron-3-nano-30b-a3b in the Python API simplest-run cell; the YAML configs elsewhere in the same notebook already used the nano model for the researcher role and note that Super has limited availability.

Flowchart

%%{init: {'theme': 'neutral'}}%%
flowchart TD
    A[User runs notebook] --> B{Execution path}
    B --> C[Python API cell\nsimplest run]
    B --> D[NAT CLI\nnat run / nat eval]
    C --> E[ChatNVIDIA\nnvidia/nemotron-3-nano-30b-a3b\nall roles: orchestrator, planner, researcher]
    D --> F[YAML config\norchestrator: gpt_oss_llm\nplanner: gpt_oss_llm\nresearcher: nemotron_nano_llm]
    E --> G[enable_thinking NOT set]
    F --> H[enable_thinking: true]
    G -.->|inconsistency| H

Loading

_{Last reviewed commit: 7e1f31b}

[AjayThorve]

LGTM