ChatGPT vs Claude vs Gemini

EvalCommunity · Practical AI Skills Lab

Test the leading AI assistants against your real Monitoring & Evaluation and International Development tasks.

7 practical tests · 8 evaluation criteria · Personalized weighted results

How to use this lab: Choose your role and goals, run the same prompts in all three assistants, score their outputs, then calculate which tool best fits your work. This lab does not connect to or automatically test the AI platforms; you run the prompts yourself and enter your observations.

STEP 1

Define your professional profile

Your professional role

Years of professional experience (optional)

Your sector or area of work Choose your three most important AI objectives

Writing reports and proposals Research and evidence synthesis Evaluation design and methodologies Qualitative data analysis Quantitative data analysis Data visualization and dashboards Automation and AI workflows Building AI agents Learning and professional development

Select up to three objectives.

Describe one recurring task you would like AI to improve (optional)

STEP 2

Prepare your three AI assistants

Open a separate conversation in each platform. Use the same model tier and available tools where possible, and record any differences in access or features.

ChatGPT

Open a new conversation and paste the setup prompt.

Open ChatGPT ↗

Claude

Open a new conversation and paste the same setup prompt.

Open Claude ↗

Gemini

Open a new conversation and paste the same setup prompt.

Open Gemini ↗

Fair-test reminder

Use identical instructions, source material and task requirements. Do not give one assistant extra context.

Master prompt: establish your M&E professional context

You are an AI assistant supporting a professional working in Monitoring and Evaluation and International Development. My professional responsibilities include evaluation design, monitoring frameworks, indicator development, qualitative and quantitative data analysis, research, and reporting. When responding: 1. Apply recognized Monitoring and Evaluation principles. 2. Use appropriate development-sector terminology. 3. Distinguish facts, assumptions, interpretations, and recommendations. 4. Explain methodological limitations. 5. Avoid inventing data, references, or evidence. 6. Provide practical examples whenever useful. 7. Structure outputs for professional use. 8. Identify where human verification is necessary. My professional role is: [INSERT YOUR ROLE]. My sector is: [INSERT YOUR SECTOR]. My main objectives are: [INSERT YOUR SELECTED OBJECTIVES]. My objective is to use AI to improve the quality, efficiency, and reliability of my professional work. Ask for clarification when essential information is missing. Use this context throughout our conversation.

STEP 3

Run the seven practical tests

Expand each test, copy its prompt, and submit it unchanged to all three assistants. Save the responses so you can compare them later.

STEP 4

Score each assistant

Rate each assistant from 1 to 5 for every criterion. Use the same standard for all three. Score only the work you actually tested.

1 = Very poor · 2 = Below expectations · 3 = Acceptable · 4 = Very good · 5 = Excellent. If you did not test a criterion, leave it blank rather than guessing.

Evaluation criterion	What to assess	ChatGPT	Claude	Gemini

STEP 5

Set your personal priorities

Assign a weight to each criterion. The weights must total 100%. Give greater weight to the capabilities that matter most in your role.

Total weight: 100%

STEP 6

Your personal AI comparison results

Enter scores and valid weights, then calculate your results. Scores are shown on a 1–5 scale.

ChatGPT

—

Awaiting scores

Claude

—

Awaiting scores

Gemini

—

Awaiting scores

Your recommendation will appear here after you calculate your results.

FINAL REFLECTION

Turn your results into an AI action plan

What will you do next?Important limitations or observations

Your scores and notes are a personal decision aid, not an independent benchmark. Verify calculations, sources, privacy safeguards and any consequential findings before using AI outputs professionally.

Keep developing your AI skills

EvalCommunity Academy

Explore practical learning opportunities in AI for Monitoring & Evaluation, International Development, and AI Agents for Evaluators.

Explore EvalCommunity Academy ↗

EvalCommunity · Monitoring | Evaluation | Learning | Development

ChatGPT vs Claude vs Gemini