Claude vs ChatGPT, Gemini & Copilot

Intermediate

What you'll learn

Understand why any 'X beats Y' scoreboard is stale within weeks
Compare assistants on the six axes that actually decide your choice
Run a fair bake-off on your own prompts instead of trusting a leaderboard
Know which skills transfer across every assistant so you're never locked in

"Which is best — Claude, ChatGPT, Gemini, or Copilot?" The honest answer: it depends on the task, and they all move fast. AILmanac is Claude-first, but we won't pretend the others don't exist — here's a fair way to decide.

Why we won't give you a scoreboard

Any "X beats Y" ranking is out of date almost immediately and rarely reflects your workload. Public benchmarks measure a task, on someone else's prompts, on the day they ran — the ranking can flip with the next model release. Instead, compare on the axes that matter to you and run your own bake-off on real tasks.

What each is generally known for

(Generalizations that shift over time — verify for your use case.)

Claude (Anthropic) — strong reasoning and coding, long-context work, a careful/steerable style, Claude Code for agentic dev, and a focus on safety. The subject of this whole site.
ChatGPT (OpenAI) — broad ecosystem, large plugin/tool surface, very widely adopted.
Gemini (Google) — deep integration with Google Workspace and Google's ecosystem.
Copilot (Microsoft/GitHub) — embedded in GitHub and Microsoft 365 where many people already work.

Compare on what matters to you

Axis	Question
Task quality	Which wins your bake-off on your prompts?
Where you work	Does it live in your editor / Office / Workspace?
Coding agent	How good is its agentic dev tool?
Privacy/compliance	Acceptable data terms for your data? (Privacy)
Cost & speed	At your volume and latency needs
Lock-in	How portable are your prompts/workflows?

The six comparison axes

Press Enter or Space to flip the card. Use the left and right arrow keys to move between cards.Term shown.

1 / 6

Run your own bake-off

A leaderboard answers someone else's question. A bake-off answers yours — and it takes an afternoon, not a research budget.

Guided walkthrough1 of 5

Pull actual prompts from your work — not toy puzzles. Include your hard cases and your boring high-volume cases; both decide the winner.

Here's a reusable harness prompt — paste it into any assistant to grade a batch of outputs against your own rubric:

A vendor-neutral bake-off grader

You are an impartial evaluator. Grade the assistant output below against my rubric.

Rubric (score each 1-5, then give a one-line reason):
- Correctness: are all facts and steps right?
- Format: does it match the requested structure exactly?
- Tone: appropriate for the audience?
- Usefulness: could I ship this as-is?

Do not reward length or confidence — only the rubric. Return a small table plus a total.

Task given to the assistant:
"""
[paste the original prompt]
"""

Assistant output to grade:
"""
[paste the output]
"""

The good news: skills transfer

Almost everything in Prompting and AI Foundations works on all of them. Learn the fundamentals here and you can wield any assistant — see Cross-AI Prompt Translation.

Check yourself

0/3

Why does this page refuse to publish a 'best assistant' scoreboard?
In a bake-off, when should you write your grading rubric?
You've invested heavily in prompts for one assistant. How locked in are you?

Key takeaways

No durable scoreboard exists — rankings flip with every release and rarely match your workload.
Decide on six axes: task quality, where you work, coding agent, privacy, cost & speed, and lock-in.
Run a real bake-off: real tasks, a rubric written first, blind scoring, and a re-run when a new model ships.
Skills transfer — learn the fundamentals once and you can wield Claude, ChatGPT, Gemini, or Copilot.

Why we won't give you a scoreboard​

What each is generally known for​

Compare on what matters to you​

Run your own bake-off​