Gemini Vision vs GPT-4o Vision: which multimodal model should you use?

import anthropic
import base64
import os
from pathlib import Path

# Using Google Cloud's Vertex AI client (recommended for Gemini)
import vertexai
from vertexai.generative_models import GenerativeModel

vertexai.init(project=os.environ["GCP_PROJECT_ID"])
model = GenerativeModel("gemini-2.5-pro")  # Gemini 2.5 Pro with vision

# Load image from disk and encode as base64
image_path = "photo.jpg"
with open(image_path, "rb") as img_file:
    image_data = base64.standard_b64encode(img_file.read()).decode("utf-8")

from vertexai.generative_models import Part

# Gemini accepts inline image data or URL: no explicit format declaration needed
response = model.generate_content([
    Part.from_data(data=image_data, mime_type="image/jpeg"),  # Native image support
    "Describe what you see in this image in 2 sentences."
])

print(response.text)

import anthropic
import base64
import os
from openai import OpenAI

client = OpenAI(api_key=os.environ["OPENAI_API_KEY"])

# Load image from disk and encode as base64 string
image_path = "photo.jpg"
with open(image_path, "rb") as img_file:
    image_data = base64.standard_b64encode(img_file.read()).decode("utf-8")

# GPT-4o requires explicit base64 encoding and mime type in message content
response = client.chat.completions.create(
    model="gpt-4o",  # GPT-4o with vision enabled
    messages=[
        {
            "role": "user",
            "content": [
                {
                    "type": "image_url",  # Must specify image_url type explicitly
                    "image_url": {
                        "url": f"data:image/jpeg;base64,{image_data}"
                    }
                },
                {
                    "type": "text",
                    "text": "Describe what you see in this image in 2 sentences."
                }
            ]
        }
    ]
)

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