diff --git a/cookbook/llm-ab-test-server/main.py b/cookbook/llm-ab-test-server/main.py
new file mode 100644
index 000000000..3102bedd2
--- /dev/null
+++ b/cookbook/llm-ab-test-server/main.py
@@ -0,0 +1,46 @@
+from flask import Flask, request, jsonify, abort, Response
+from flask_cors import CORS
+from litellm import completion
+import os, dotenv
+import random
+dotenv.load_dotenv()
+
+# TODO: set your keys in .env or here:
+# os.environ["OPENAI_API_KEY"] = "" # set your openai key here or in your .env
+# see supported models, keys here:
+
+
+app = Flask(__name__)
+CORS(app)
+
+@app.route('/')
+def index():
+ return 'received!', 200
+
+# Dictionary of LLM functions with their A/B test ratios, should sum to 1 :)
+llm_dict = {
+ "gpt-4": 0.2,
+ "together_ai/togethercomputer/llama-2-70b-chat": 0.4,
+ "claude-2": 0.2,
+ "claude-1.2": 0.2
+}
+
+
+@app.route('/chat/completions', methods=["POST"])
+def api_completion():
+ data = request.json
+ try:
+ # pass in data to completion function, unpack data
+ selected_llm = random.choices(list(llm_dict.keys()), weights=list(llm_dict.values()))[0]
+ data['model'] = selected_llm
+ response = completion(**data)
+ except Exception as e:
+ print(f"got error{e}")
+ return response, 200
+
+
+if __name__ == "__main__":
+ from waitress import serve
+ print("starting server")
+ serve(app, host="0.0.0.0", port=5000, threads=500)
+
diff --git a/cookbook/llm-ab-test-server/readme.md b/cookbook/llm-ab-test-server/readme.md
new file mode 100644
index 000000000..2c7afd6bd
--- /dev/null
+++ b/cookbook/llm-ab-test-server/readme.md
@@ -0,0 +1,97 @@
+
+ 🚅 LiteLLM - A/B Testing LLMs in Production
+
+
+
Call all LLM APIs using the OpenAI format [Anthropic, Huggingface, Cohere, Azure OpenAI etc.]
+
+
+
+
+
+
+LiteLLM allows you to call 100+ LLMs using completion
+This template server allows you to define LLMs with their A/B test ratios
+
+```python
+llm_dict = {
+ "gpt-4": 0.2,
+ "together_ai/togethercomputer/llama-2-70b-chat": 0.4,
+ "claude-2": 0.2,
+ "claude-1.2": 0.2
+}
+```
+
+Litellm simplifies I/O with all models, the server simply makes a `litellm.completion()` call to the selected model
+
+
+
+- Translating inputs to the provider's completion and embedding endpoints
+- Guarantees [consistent output](https://litellm.readthedocs.io/en/latest/output/), text responses will always be available at `['choices'][0]['message']['content']`
+- Exception mapping - common exceptions across providers are mapped to the [OpenAI exception types](https://help.openai.com/en/articles/6897213-openai-library-error-types-guidance)
+# Usage
+
+
+ 
+
+
+
+```
+pip install litellm
+```
+
+```python
+from litellm import completion
+
+## set ENV variables
+os.environ["OPENAI_API_KEY"] = "openai key"
+os.environ["COHERE_API_KEY"] = "cohere key"
+os.environ["ANTHROPIC_API_KEY"] = "anthropic key"
+
+messages = [{ "content": "Hello, how are you?","role": "user"}]
+
+# openai call
+response = completion(model="gpt-3.5-turbo", messages=messages)
+
+# cohere call
+response = completion(model="command-nightly", messages=messages)
+
+# anthropic
+response = completion(model="claude-2", messages=messages)
+```
+
+Stable version
+```
+pip install litellm==0.1.424
+```
+
+
+
+
+
+
+# support / talk with founders
+- [Schedule Demo 👋](https://calendly.com/d/4mp-gd3-k5k/berriai-1-1-onboarding-litellm-hosted-version)
+- [Community Discord 💭](https://discord.gg/wuPM9dRgDw)
+- Our numbers 📞 +1 (770) 8783-106 / +1 (412) 618-6238
+- Our emails ✉️ ishaan@berri.ai / krrish@berri.ai
+
+# why did we build this
+- **Need for simplicity**: Our code started to get extremely complicated managing & translating calls between Azure, OpenAI, Cohere
+
+
+ 
+
+
+ 
+
+ 
+
+