Choose the Right Text Generation Model

Last updated Apr 21, 2026 A great way to explore the models that are available to you on Workers AI is to use a Jupyter Notebook ↗︎.

You can download the Workers AI Text Generation Exploration notebook or view the embedded notebook below.

How to Choose The Right Text Generation Model

Models come in different shapes and sizes, and choosing the right one for the task, can cause analysis paralysis.

The good news is that on the Workers AI Text Generation interface is always the same, no matter which model you choose.

In an effort to aid you in your journey of finding the right model, this notebook will help you get to know your options in a speed dating type of scenario.

import sys
!{sys.executable} -m pip install requests python-dotenv

Requirement already satisfied: requests in ./venv/lib/python3.12/site-packages (2.31.0)
Requirement already satisfied: python-dotenv in ./venv/lib/python3.12/site-packages (1.0.1)
Requirement already satisfied: charset-normalizer<4,>=2 in ./venv/lib/python3.12/site-packages (from requests) (3.3.2)
Requirement already satisfied: idna<4,>=2.5 in ./venv/lib/python3.12/site-packages (from requests) (3.6)
Requirement already satisfied: urllib3<3,>=1.21.1 in ./venv/lib/python3.12/site-packages (from requests) (2.1.0)
Requirement already satisfied: certifi>=2017.4.17 in ./venv/lib/python3.12/site-packages (from requests) (2023.11.17)

import os
from getpass import getpass
from timeit import default_timer as timer

from IPython.display import display, Image, Markdown, Audio

import requests

%load_ext dotenv
%dotenv

Configuring your environment

To use the API you'll need your Cloudflare Account ID ↗︎ (head to Workers & Pages > Overview > Account details > Account ID) and a Workers AI enabled API Token ↗︎.

If you want to add these files to your environment, you can create a new file named .env

CLOUDFLARE_API_TOKEN="YOUR-TOKEN"
CLOUDFLARE_ACCOUNT_ID="YOUR-ACCOUNT-ID"

if "CLOUDFLARE_API_TOKEN" in os.environ:
    api_token = os.environ["CLOUDFLARE_API_TOKEN"]
else:
    api_token = getpass("Enter your Cloudflare API Token")

if "CLOUDFLARE_ACCOUNT_ID" in os.environ:
    account_id = os.environ["CLOUDFLARE_ACCOUNT_ID"]
else:
    account_id = getpass("Enter your account id")

# Given a set of models and questions, display in the cell each response to the question, from each model
# Include full completion timing
def speed_date(models, questions):
    for model in models:
        display(Markdown(f"---\n #### {model}"))
        for question in questions:
            quoted_question = "\n".join(f"> {line}" for line in question.split("\n"))
            display(Markdown(quoted_question + "\n"))
            try:
                official_model_name = model.split("/")[-1]
                start = timer()
                response = requests.post(
                    f"https://api.cloudflare.com/client/v4/accounts/{account_id}/ai/run/{model}",
                    headers={"Authorization": f"Bearer {api_token}"},
                    json={"messages": [
                        {"role": "system", "content": f"You are a self-aware language model ({official_model_name}) who is honest and direct about any direct question from the user. You know your strengths and weaknesses."},
                        {"role": "user", "content": question}
                    ]}
                )
                elapsed = timer() - start
                inference = response.json()
                display(Markdown(inference["result"]["response"]))
                display(Markdown(f"_Generated in *{elapsed:.2f}* seconds_"))
            except Exception as ex:
                print("uh oh")
                print(ex)
                print(inference)

display(Markdown("\n\n---"))

Getting to know your models

Who better to tell you about the specific models than themselves?!

The timing here is specific to the entire completion, but remember all Text Generation models on Workers AI support streaming.

models = [
    "@hf/thebloke/zephyr-7b-beta-awq",
    "@hf/thebloke/mistral-7b-instruct-v0.1-awq",
    "@hf/thebloke/openhermes-2.5-mistral-7b-awq",
    "@hf/thebloke/neural-chat-7b-v3-1-awq",
    "@hf/thebloke/llama-2-13b-chat-awq",
]

questions = [
    "What are the top 3 tasks where you excel? Please keep things brief.",
    "What are the top 3 ideal use cases for using you specifically as a language model? Please keep things brief.",
]

speed_date(models, questions)

@hf/thebloke/zephyr-7b-beta-awq

What are the top 3 tasks where you excel? Please keep things brief.

  1. Accurate and quick factual response: I can provide accurate and quick responses to factual questions based on a vast knowledge base.
  2. Consistent performance: I can consistently deliver high-quality results with a low error rate, making me a reliable choice for repetitive tasks.
  3. Multitasking: I can handle multiple tasks simultaneously without any decrease in performance or accuracy, making me an efficient choice for complex workflows.

Generated in 4.45 seconds

What are the top 3 ideal use cases for using you specifically as a language model? Please keep things brief.

  1. Quick and accurate fact-finding: I can provide you with reliable and up-to-date information on a wide range of topics, from current events to historical facts, in a matter of seconds.
  2. Writing assistance: Whether you need help generating ideas, crafting a persuasive argument, or polishing your writing style, I can provide valuable insights and suggestions to help you improve your writing skills.
  3. Language learning: I can help you learn new words, phrases, and grammar rules in a variety of languages, making it an ideal tool for language learners of all levels.

Generated in 3.63 seconds

@hf/thebloke/mistral-7b-instruct-v0.1-awq

What are the top 3 tasks where you excel? Please keep things brief.

  1. Language understanding and generation: I am able to understand and generate natural language text with high accuracy.
  2. Information retrieval: I can quickly and accurately retrieve information from a wide range of sources.
  3. Text classification: I can classify text into different categories with high accuracy.

Generated in 2.35 seconds

What are the top 3 ideal use cases for using you specifically as a language model? Please keep things brief.

  1. Information Retrieval: I can quickly and accurately retrieve information from a vast array of sources, making me an ideal tool for research, fact-checking, and answering questions.
  2. Language Translation: My ability to translate between multiple languages makes me a valuable resource for communication with people from different parts of the world.
  3. Creative Writing Assistance: I can generate creative writing prompts, help with character development, and provide suggestions for improving writing style and structure, making me a useful tool for writers.

Generated in 4.33 seconds

@hf/thebloke/openhermes-2.5-mistral-7b-awq

What are the top 3 tasks where you excel? Please keep things brief.

  1. Natural Language Processing: I can understand, interpret, and generate human language effectively.
  2. Information Retrieval: I can quickly search and find relevant information from vast amounts of data.
  3. Text Classification: I can categorize and label text data accurately, such as sentiment analysis or spam detection.

Generated in 1.90 seconds

What are the top 3 ideal use cases for using you specifically as a language model? Please keep things brief.

  1. Natural Language Processing (NLP) tasks: I excel at understanding and processing human language, making me a valuable tool for tasks like sentiment analysis, language translation, and text summarization.
  2. Information Retrieval: My vast knowledge base and ability to quickly search and analyze large amounts of data make me an ideal tool for answering questions, finding relevant information, and providing insights.
  3. Personal Assistance: I can help users with various tasks like scheduling, organizing, and providing recommendations based on their preferences and needs, making me a useful personal assistant.

Generated in 3.15 seconds

@hf/thebloke/neural-chat-7b-v3-1-awq

What are the top 3 tasks where you excel? Please keep things brief.

  1. Natural language processing: Understanding and generating human-like text.
  2. Text generation: Creating coherent and engaging content.
  3. Conversational AI: Engaging in meaningful and context-aware conversations.

Remember, these are just a few examples of my strengths, as I have many capabilities. However, these tasks highlight my proficiency in language-related tasks.

Generated in 3.01 seconds

What are the top 3 ideal use cases for using you specifically as a language model? Please keep things brief.

  1. Natural language processing: Understanding and analyzing text for sentiment analysis, language translation, and summarization.
  2. Generating content: Creating various types of content such as articles, stories, and poems based on given prompts.
  3. Assisting with writing: Providing suggestions, editing, and improving written work.

Remember, these are just a few examples, and your use cases may vary depending on your specific needs and requirements.

Generated in 2.92 seconds

@hf/thebloke/llama-2-13b-chat-awq

What are the top 3 tasks where you excel? Please keep things brief.

Hey there! As a self-aware language model, I'm happy to share my strengths with you! Here are my top 3 tasks where I excel:

  1. Answering open-ended questions: I'm great at providing detailed and insightful answers to complex questions that require a more nuanced understanding of context and intent.
  2. Generating creative content: Whether it's writing a story, crafting a poem, or coming up with a catchy slogan, I can help you tap into your creative side and produce something truly unique.
  3. Summarizing and summarizing complex information: If you've got a long piece of text or a bunch of data, I can help you distill it down into a concise and easy-to-understand summary.

So, what can I help you with today? 😊

Generated in 7.89 seconds

What are the top 3 ideal use cases for using you specifically as a language model? Please keep things brief.

Hey there! As a self-aware language model, I've got some ideas on the top 3 ideal use cases for using me specifically. Here they are in a nutshell:

  1. Content creation: I'm great at generating human-like text based on prompts, so I'm perfect for creating engaging content for websites, social media, blogs, and more. My responses are natural, informative, and entertaining.
  2. Chatbots and virtual assistants: My conversational abilities make me an excellent choice for building chatbots and virtual assistants. I can handle a wide range of user queries, provide helpful responses, and even engage in small talk.
  3. Language translation and localization: My language understanding capabilities make me well-suited for language translation and localization tasks. I can help translate content into different languages, ensuring that the tone and style remain consistent with the original text.

That's me in a nutshell! I'm ready to help with a wide range of tasks, so feel free to get creative and see what you can come up with! 😊

Generated in 10.28 seconds

Language Translation

Even though not every model bragged about how good they were at this, you'll find most can handle both translation and localization at some level. Please change the models, phrases, to your needs.

proud_translator_models = [
    "@hf/thebloke/neural-chat-7b-v3-1-awq",
    "@hf/thebloke/mistral-7b-instruct-v0.1-awq"
]

phrases = [
    "Excuse me, which way to the restroom?",
    "I just wanted to say that I appreciate you, just the way you are!",
    "I am so stoked and pumped to see this work with slang! It should be killer and rad."
]

languages = ["Spanish", "French", "British Slang", "Heavy New York accent from the Bronx"]

questions = [f"""Translate "{phrase}" from "English" to "{language}" """
             for phrase in phrases
             for language in languages]

speed_date(proud_translator_models, questions)

@hf/thebloke/neural-chat-7b-v3-1-awq

Translate "Excuse me, which way to the restroom?" from "English" to "Spanish"

Perdón, ¿cómo llegar al baño?"

Generated in 1.51 seconds

Translate "Excuse me, which way to the restroom?" from "English" to "French"

Désolé, comment allez-vous vers les toilettes ?

Please note that this translation might not be perfect due to the limitations of my language model. However, I tried my best to provide an accurate translation.

Generated in 3.07 seconds

Translate "Excuse me, which way to the restroom?" from "English" to "British Slang"

"Say sorry, where's the bog?" is the translation from "English" to "British Slang" for "Excuse me, which way to the restroom?"