Deprecated Patterns

Deprecated Patterns

Common mistakes Claude makes when generating Gemini API code. Check this BEFORE writing any Gemini code.

Source: curated from ~/Documents/google-gemini-context/CLAUDE.md

Package Names

Imports

JavaScript

// WRONG
import { GoogleGenerativeAI } from "@google/generative-ai";

// CORRECT
import { GoogleGenAI } from "@google/genai";

Python

# WRONG
import google.generativeai as genai
genai.configure(api_key=os.environ["GEMINI_API_KEY"])

# CORRECT
from google import genai
client = genai.Client()  # auto-reads GEMINI_API_KEY

Initialisation

JavaScript

// WRONG - old pattern
const genAI = new GoogleGenerativeAI(process.env.GEMINI_API_KEY);
const model = genAI.getGenerativeModel({ model: "gemini-pro" });
const result = await model.generateContent("prompt");

// CORRECT - new pattern
const ai = new GoogleGenAI({});  // auto-reads GEMINI_API_KEY
const response = await ai.models.generateContent({
  model: "gemini-2.5-flash",
  contents: "prompt"
});

Python

# WRONG - old pattern
genai.configure(api_key=os.environ["GEMINI_API_KEY"])
model = genai.GenerativeModel("gemini-pro")
response = model.generate_content("prompt")

# CORRECT - new pattern
client = genai.Client()  # auto-reads GEMINI_API_KEY
response = client.models.generate_content(
    model="gemini-2.5-flash",
    contents="prompt"
)

Model IDs

Default recommendation: gemini-2.5-flash for most tasks, gemini-2.5-pro for complex reasoning.

Chat API

JavaScript

// WRONG - old pattern
const chat = model.startChat();
const result = await chat.sendMessage("Hello");

// CORRECT - new pattern
const chat = await ai.chats.create({ model: "gemini-2.5-flash" });
const result = await chat.send("Hello");

Python

# WRONG - old streaming parameter
response = model.generate_content("prompt", stream=True)

# CORRECT - config-based streaming
response = client.models.generate_content(
    model="gemini-2.5-flash",
    contents="prompt",
    config={"stream": True}
)

Safety Categories

Claude consistently forgets the 5th category and uses the wrong name for the 4th:

# WRONG - 4 categories, wrong name
categories = [
    "HARM_CATEGORY_HARASSMENT",
    "HARM_CATEGORY_HATE_SPEECH",
    "HARM_CATEGORY_SEXUALLY_EXPLICIT",
    "HARM_CATEGORY_DANGEROUS_CONTENT"       # wrong: no _CONTENT
]

# CORRECT - 5 categories
categories = [
    "HARM_CATEGORY_HARASSMENT",
    "HARM_CATEGORY_HATE_SPEECH",
    "HARM_CATEGORY_SEXUALLY_EXPLICIT",
    "HARM_CATEGORY_DANGEROUS",               # no _CONTENT suffix
    "HARM_CATEGORY_CIVIC_INTEGRITY"          # don't forget this one
]

REST API Headers

# WRONG - Claude often capitalises this
-H "X-Goog-Api-Key: $GEMINI_API_KEY"

# CORRECT - lowercase
-H "x-goog-api-key: $GEMINI_API_KEY"

Parameter Names

Rate Limits

• There are NO daily limits -- only per-minute (RPM, TPM)
• Claude sometimes invents "1,500 requests per day" -- this is wrong
• Free tier: per-model RPM/TPM limits only