Calculator
Migration guides

The Switchboard

Turn price savings into action. Step-by-step guides with code snippets, compatibility scores and gotchas.

Prices checked 13 Nov 2026

Guides

OpenAI to Anthropic

Compatibility
85%
Difficulty
Easy
Time
1-2 hours
Cheapest output price
Same priceGPT-6 Luna $0.50 vs Claude Haiku 5.5 $0.50 per million output tokens

Migration steps

  1. Install the Anthropic SDK: npm install @anthropic-ai/sdk
  2. Replace API key environment variable: OPENAI_API_KEY → ANTHROPIC_API_KEY
  3. Update the client initialization code
  4. Modify message format (role/content structure is similar)
  5. Update function calling to use Claude's tool_use format
  6. Test with a subset of your prompts to verify output quality
Before (OpenAI)
import OpenAI from 'openai';

const client = new OpenAI({
  apiKey: process.env.OPENAI_API_KEY,
});

const response = await client.chat.completions.create({
  model: "gpt-5.4",
  messages: [
    { role: "system", content: "You are helpful." },
    { role: "user", content: "Hello!" }
  ],
});

console.log(response.choices[0].message.content);
After (Anthropic)
import Anthropic from '@anthropic-ai/sdk';

const client = new Anthropic({
  apiKey: process.env.ANTHROPIC_API_KEY,
});

const response = await client.messages.create({
  model: "claude-sonnet-5-5",
  max_tokens: 1024,
  system: "You are helpful.",
  messages: [
    { role: "user", content: "Hello!" }
  ],
});

// Thinking is on by default, so content can start with a thinking block
for (const block of response.content) {
  if (block.type === "text") console.log(block.text);
}

Watch out for

  • System prompts work differently -- Claude uses a dedicated system parameter
  • Function calling syntax differs -- Claude uses tool_use blocks
  • Token counting methods are different
  • Max context window varies by model

OpenAI to Google

Compatibility
75%
Difficulty
Medium
Time
2-4 hours
Cheapest output price
▼ 20% lowerGPT-6 Luna $0.50 vs Gemini 2.5 Flash-Lite $0.40 per million output tokens

Migration steps

  1. Install Google AI SDK: npm install @google/generative-ai
  2. Replace API key: OPENAI_API_KEY → GOOGLE_API_KEY
  3. Update client initialization with GoogleGenerativeAI
  4. Convert messages to Gemini's parts-based format
  5. Update streaming handlers if using SSE
  6. Configure safety settings as needed
Before (OpenAI)
import OpenAI from 'openai';

const client = new OpenAI({
  apiKey: process.env.OPENAI_API_KEY,
});

const response = await client.chat.completions.create({
  model: "gpt-5.4",
  messages: [
    { role: "user", content: "Hello!" }
  ],
});
After (Google)
import { GoogleGenerativeAI } from '@google/generative-ai';

const genAI = new GoogleGenerativeAI(
  process.env.GOOGLE_API_KEY
);

const model = genAI.getGenerativeModel({
  model: "gemini-3.8-flash"
});

const result = await model.generateContent("Hello!");
console.log(result.response.text());

Watch out for

  • Gemini uses a different message format (parts-based)
  • Safety settings are more granular in Gemini
  • Streaming API differs significantly
  • Function calling uses a different schema format

OpenAI to DeepSeek

Compatibility
95%
Difficulty
Easy
Time
30 minutes
Cheapest output price
▲ 140% higherGPT-6 Luna $0.50 vs DeepSeek V4.1 Flash $1.20 per million output tokens

Migration steps

  1. Update the base URL to DeepSeek's endpoint
  2. Replace API key: OPENAI_API_KEY → DEEPSEEK_API_KEY
  3. Change model name to deepseek-flash (or deepseek-v4-pro)
  4. Test -- most code works without changes
Before (OpenAI)
import OpenAI from 'openai';

const client = new OpenAI({
  apiKey: process.env.OPENAI_API_KEY,
});

const response = await client.chat.completions.create({
  model: "gpt-5.4",
  messages: [
    { role: "user", content: "Hello!" }
  ],
});
After (DeepSeek)
import OpenAI from 'openai';

const client = new OpenAI({
  apiKey: process.env.DEEPSEEK_API_KEY,
  baseURL: "https://api.deepseek.com",
});

const response = await client.chat.completions.create({
  model: "deepseek-flash",
  messages: [
    { role: "user", content: "Hello!" }
  ],
});
// Same response format -- it just works!

Watch out for

  • DeepSeek uses OpenAI-compatible API -- minimal changes needed
  • Vision works on deepseek-flash but not on deepseek-v4-pro
  • Weekday peak hours (01:00-04:00 and 06:00-10:00 UTC) bill at double the off-peak rate
  • Rate limits and quotas differ
  • Response format may have subtle differences

Anthropic to OpenAI

Compatibility
85%
Difficulty
Easy
Time
1-2 hours
Cheapest output price
Same priceClaude Haiku 5.5 $0.50 vs GPT-6 Luna $0.50 per million output tokens

Migration steps

  1. Install OpenAI SDK: npm install openai
  2. Replace API key: ANTHROPIC_API_KEY → OPENAI_API_KEY
  3. Move system prompt into messages array
  4. Update response parsing from text blocks in content to choices[0].message.content
  5. Update tool calling format if used
Before (Anthropic)
import Anthropic from '@anthropic-ai/sdk';

const client = new Anthropic();

const response = await client.messages.create({
  model: "claude-sonnet-5-5",
  max_tokens: 1024,
  system: "You are helpful.",
  messages: [
    { role: "user", content: "Hello!" }
  ],
});
After (OpenAI)
import OpenAI from 'openai';

const client = new OpenAI();

const response = await client.chat.completions.create({
  model: "gpt-5.4",
  messages: [
    { role: "system", content: "You are helpful." },
    { role: "user", content: "Hello!" }
  ],
});

Watch out for

  • OpenAI uses system messages in the messages array, not a separate parameter
  • Function calling syntax is different (tools vs tool_use)
  • Response structure differs (choices[] vs content[])

Next steps

OpenAI pricingEvery OpenAI model and its price.Anthropic pricingEvery Anthropic model and its price.Google pricingEvery Google model and its price.DeepSeek pricingEvery DeepSeek model and its price.Cost calculatorEstimate what the switch would save on your workload.