When google grounded search confidence search are missing?

Google Workspace Developers Q&A
1

Hi,

I really like the google grounded search recent feature, I have a use case where I will benefit a lot from the grounded search confidence scores.

Yesterday(16.06.25), the same request I made today to google grounding search was returning confidence scores, but today it isn’t:

grounding support result: confidence_scores=None, grounding_chunk_indices=[0, 1] segment=Segment(end_index=189, part_index=None, start_index=82, text=‘…’)

I have not changed the model, namely “gemini-2.0-flash”. This is a sample code for reference :
from google import genai
from google.genai.types import (
Tool,
GenerateContentConfig,
GoogleSearch,
)

client = genai.Client()
model_id = “gemini-2.0-flash”
google_search_tool = Tool(
google_search = GoogleSearch()
)
response = client.models.generate_content(
model=model_id,
contents=“”“As of June 2025, summarize the BVNK Latest 20 news in press as of May and June 2025 {“summary” : “Summary of the last 20 news in the press””",
config=GenerateContentConfig(
tools=[google_search_tool],

response_modalities=[“TEXT”],

)
)

for each in response.candidates[0].content.parts:
print(each.text)

Is there some specific model or smth I should change?

I will really benefit from having the confidence scores.

Thank you in advance,
Stanka

2

Until Google formally documents confidence_scores as a stable part of the response structure (or provides toggles), their presence is subject to change, and unfortunately, you as a developer don’t have full control over that part of the pipeline.

If these scores are mission-critical for your use case, maybe you can:
Log all available metadata in the response so you can spot when fields are added/removed
File a support ticket or feedback report with Google if you have access to an enterprise account (this may nudge them to provide more stable guarantees).