What are the Gemini API's multimodal input capabilities?

from google import genai
from pathlib import Path

client = genai.Client()

# Upload a file via the Files API (for large files)
my_video = client.files.upload(file=Path("presentation.mp4"))

# Multimodal request: video + text question
interaction = client.interactions.create(
    model="gemini-3.5-flash",
    input=[
        {"type": "text", "text": "Summarise the key points from this video."},
        {"type": "video", "uri": my_video.uri, "mime_type": "video/mp4"}
    ]
)
print(interaction.output_text)

# Image inline (base64 for small images, Files API for large):
import base64
image_data = base64.b64encode(Path("diagram.png").read_bytes()).decode()
interaction = client.interactions.create(
    model="gemini-3.5-flash",
    input=[
        {"type": "text", "text": "Explain this architecture diagram."},
        {"type": "image", "data": image_data, "mime_type": "image/png"}
    ]
)