What is the observable execution steps feature in the Gemini Interactions API?

from google import genai
client = genai.Client()

interaction = client.interactions.create(
    model="gemini-3.5-flash",
    input="What is the square root of the number of stars in the Milky Way?",
    tools=[{"google_search": {}}],
)

# Inspect every step the model took:
for i, step in enumerate(interaction.steps):
    print(f"Step {i}: type={step.type}")

    if step.type == "thought":
        # Encrypted thinking (content not shown but signature confirms it happened)
        print(f"  Thinking... (signature: {step.signature[:20]}...)")

    elif step.type == "tool_call":
        print(f"  Tool call: {step.tool_call.name}")
        print(f"  Arguments: {step.tool_call.arguments}")

    elif step.type == "tool_result":
        print(f"  Tool result received")

    elif step.type == "code_execution":
        print(f"  Code: {step.code}")
        print(f"  Output: {step.output}")

    elif step.type == "model_output":
        print(f"  Final answer: {step.content[0].text}")

# Use steps to build a UI showing real-time reasoning:
for step in interaction.steps:
    if step.type == "tool_call":
        show_ui("Searching the web for: " + step.tool_call.arguments.get("query",""))
    elif step.type == "thought":
        show_ui("Thinking...")
    elif step.type == "model_output":
        show_ui("Answer: " + step.content[0].text)