Step 01
Define Your Benchmark Tasks
Identify 3-7 core tasks you perform regularly, such as drafting documents, coding, summarizing notes, or triaging your inbox.
IntermediatePersonal
How to Conduct a Blind Taste Test to Find the Best AI Model for Your Work
Systematically evaluate AI models like GPT-6 and Opus 5.5 to find the best fit for your tasks. This blind test method helps you choose tools based on real-world performance and personal preference, not just industry benchmarks.
Try this with your AI assistant
Download the skill for this workflow. Version 1.0.0.
Before you start
Tools used
- CodexVisitCodex
OpenAI's cloud-based AI software engineering agent that can execute code, run tests, and handle complex multi-file tasks autonomously.
- ClaudeVisitClaude
Anthropic AI assistant
The workflow
Follow the sequence once, then adapt the prompts, checks, and handoffs to your own setup.
6 steps
Step 02
Set Up a Blind Test Environment
Use a tool or ask a colleague to run your prompts on different models and label the outputs anonymously (e.g., Model A, Model B) so you can't tell which is which.
Step 03
Run Identical Prompts Across Models
Submit the exact same prompt for each of your benchmark tasks to every model you are testing. This ensures a fair and direct comparison of their outputs.
Step 04
Score the Anonymous Outputs
Review each output without knowing the source model. Rate them on a personal 'vibe' scale (e.g., 1-5) and add notes on what you liked or disliked about the style, structure, and quality.
Step 05
Reveal and Analyze the Results
Un-blind the model names and aggregate your scores in a spreadsheet. Calculate average scores and identify which models performed best on average and on specific tasks.
Step 06
Select Your Go-To Models
Based on your analysis, assign specific models to specific tasks in your daily workflow, optimizing for both quality and cost to build your personal AI toolkit.
Build your next product with ChatPRD
Turn an idea into a PRD, user stories, and a plan.
Related workflows
intermediateProductCodex
Create an Immersive AI Language Tutor
Build a voice-first language practice app that forces immersion by refusing to switch to the user's native language. The AI tutor provides corrections in real-time, within the flow of the conversation, to simulate an authentic learning environment.
Open
advancedOperationsChatGPTCodex+1
How to Manage Business Operations with Role-Based AI Agents
Organize your business operations by creating a 'fellowship' of specialized AI agents. This workflow shows how to silo responsibilities into projects and assign distinct roles to AI assistants to automate administrative tasks and context retrieval.
Open
intermediateProductCodex
Build a Native Mobile App with AI and Clear Product Direction
Use Codex to help build a native mobile app, then guide the design and review the resulting code. William brought web-development and product experience to the project.
Open
Start shipping better products.
Join 250,000+ product managers who use ChatPRD to write better docs, align teams faster, and build products users love.
Free to start · 14-day Pro trial · SOC 2 Type II
