[Hugo-DDT]

Please ensure you have read the contribution guide before creating a pull request.

Link to Issue or Description of Change

1. Link to an existing issue (if applicable):

• Fixes: gen_ai.choice log reports gen_ai.system=gemini for OpenAI and Groq models through LiteLlm #7631

2. Or, if no issue exists, describe the change:

Problem:
With ADK_TELEMETRY_SCHEMA_VERSION_OPT_IN=1, gen_ai.choice logs report gen_ai.system=gemini for LiteLLM models such as openai/gpt-4o and groq/llama-3.1-8b-instant, while the span and system/user message logs correctly report openai / groq. Anthropic is unaffected because bare claude-* ids already resolve via the Anthropic pattern.

Root cause: _inference_system_name preferred llm_response.model_version. LiteLLM returns that without a provider prefix (e.g. gpt-4o), so _resolve_gen_ai_system_name falls back to Gemini.

Solution:
When model_version does not itself identify a gen_ai.system provider (no Gemini match, no Anthropic pattern, no <provider>/<model> prefix), fall back to the agent-configured request model id (the same id the span already uses). Added _model_identifies_gen_ai_system to share that check with the resolution rules.

Testing Plan

Unit Tests:

• I have added or updated unit tests for my change.
• All unit tests pass locally.

New test_choice_log_system_name_when_model_version_lacks_provider covers openai/gpt-4o, groq/llama-3.1-8b-instant, anthropic, and gemini when model_version is a bare id.

pytest tests/unittests/telemetry/test_spans.py::test_choice_log_system_name_when_model_version_lacks_provider \
       tests/unittests/telemetry/test_spans.py::test_generate_content_span_system_name_follows_model \
       tests/unittests/telemetry/test_spans.py::test_resolve_gen_ai_system_name \
       tests/unittests/telemetry/test_spans.py::test_resolve_gen_ai_system_name_never_names_a_path_segment
# 24 passed

pytest tests/unittests/telemetry/test_spans.py tests/unittests/telemetry/test_stable_semconv.py
# 190 passed

Red/green: the new openai/groq cases failed on main (gemini vs expected provider) and pass with the fix.

Manual End-to-End (E2E) Tests:

Not run (repro in #7631 uses LiteLLM mock_response; covered by the unit regression above).

Checklist

• I have read the CONTRIBUTING.md document.
• I have performed a self-review of my own code.
• I have commented my code, particularly in hard-to-understand areas.
• I have added tests that prove my fix is effective or that my feature works.
• New and existing unit tests pass locally with my changes.
• I have manually tested my changes end-to-end.
• Any dependent changes have been merged and published in downstream modules.

Additional context

N/A