What problem do you want to solve?

Add support for collecting additional token usage metadata from Google GenAI responses:

• usage_metadata.thoughts_token_count (Gemini 2.5 thinking tokens)
• usage_metadata.cached_content_token_count

Describe the solution you'd like

1. Collect thoughts_token_count as gcp.gen_ai.usage.thoughts_token_count
   • Using GCP vendor prefix since not in OTEL standard
2. Collect cached_content_token_count as gen_ai.usage.cache_read.input_tokens
   • Using existing OTEL standard attribute

Describe alternatives you've considered

No response

Additional Context

Motivation

• Gemini 2.5 models expose thinking tokens for extended reasoning capabilities
• Cached tokens are important for cost tracking and performance analysis
• Users need visibility into all token types for accurate billing and monitoring

Current State

Currently, only prompt_token_count and candidates_token_count are collected:
https://github.com/open-telemetry/opentelemetry-python-contrib/blob/main/instrumentation-genai/opentelemetry-instrumentation-google-genai/src/opentelemetry/instrumentation/google_genai/generate_content.py#L425-L435

I'm willing to submit a PR for this feature. 👍🏻

Would you like to implement a fix?

Yes

Tip

_{React with 👍 to help prioritize this issue. Please use comments to provide useful context, avoiding +1 or me too, to help us triage it. Learn more here.}