[Bug]: `litellm.utils.supports_pdf_input` returns `False` for gemini 2.5 models

[bmanczak]

What happened?

litellm.utils.supports_pdf_input("gemini/gemini-2.5-flash") returns False, but Gemini 2.5 Flash actually supports PDF input (at least locally).

Code to reproduce

#!/usr/bin/env python3
import base64
from pathlib import Path

from litellm import completion
from litellm.utils import supports_pdf_input

print(f"litellm.utils.supports_pdf_input('gemini/gemini-2.5-flash') = {supports_pdf_input('gemini/gemini-2.5-flash')}")

# Read and encode PDF
pdf_path = "/workspace/tests/multi_modal_test_assets/cover_book.pdf"
pdf_b64 = base64.b64encode(Path(pdf_path).read_bytes()).decode()

# API call
response = completion(
    model="gemini/gemini-2.5-flash",
    messages=[
        {
            "role": "user",
            "content": [
                {"type": "text", "text": "What is this book about? Give a very brief summary."},
                {"type": "file", "file": {"file_data": f"data:application/pdf;base64,{pdf_b64}"}},
            ],
        }
    ],
    temperature=0.0,
    max_tokens=150,
)

print(response.choices[0].message.content)

Output:

litellm.utils.supports_pdf_input('gemini/gemini-2.5-flash') = False
This book is **"The Little Prince" by Antoine de

Are you a ML Ops Team?

No

What LiteLLM version are you on ?

1.72.4

Twitter / LinkedIn details

No response