Edapt HSBR Works

Created by Hisham Basheer (HSBR)

Model comparison
Back
Question Models Comparison Verdicts Reflection HSBR Note

Task 7: ChatGPT Model Comparison — Understanding AI Intelligence Levels

Compare how three ChatGPT models answer the same work question.

3 Models
Explain step by step how to use Git with GitHub for a small project that has 5 or more developers. Cover everything from setting up the repository, cloning, branching strategy, commit and push rules, pull requests, code reviews, resolving conflicts, and merging to main. Also explain how to keep the project organized, avoid merge conflicts, and follow best practices for collaboration. Include concrete examples of Git commands and GitHub workflows that a small dev team should follow.
1
ChatGPT 4.0 (Standard)
Baseline

Clear tutorial style with essential commands and a simple GitHub Flow.

  • Depth: Medium — covers basics well.
  • Structure: Checklist/cheat-sheet.
  • Evidence: None; common practices.
  • Reasoning: Procedural; minimal trade-offs.
  • Actionability: Good for onboarding.
Most Practical
2
ChatGPT o3 (Advanced reasoning)
Reasoner

Opinionated playbook (branch protection, CODEOWNERS, CI gates), with concise rationale.

  • Depth: High — includes governance & CI.
  • Structure: Playbook with rules & examples.
  • Evidence: Best-practice assertions.
  • Reasoning: Weighs rebase vs merge, squash, etc.
  • Actionability: Excellent (ready to adopt).
Most Thoughtful
3
ChatGPT Deep Research
Research

Handbook style with trade-offs and (if desired) citations to docs.

  • Depth: Very high — rationale & risks.
  • Structure: Handbook/guide.
  • Evidence: Can include source links.
  • Reasoning: Strong “why” behind choices.
  • Actionability: Good but verbose.

Model Comparison

Scroll on mobile
Dimension ChatGPT 4.0 (Standard) ChatGPT o3 (Advanced reasoning) ChatGPT Deep Research
Depth of analysis Medium — solid basics High — includes governance & CI Very high — trade-offs & risks
Response structure Simple tutorial / checklist Playbook with rules & examples Handbook style, comprehensive
Evidence / examples Generic commands Concrete commands + config snippets Examples + optional citations
Reasoning process Direct steps Explains trade-offs (rebase vs merge, squash) Deep rationale; risks & alternatives
Actionability Good for onboarding Excellent — copy-paste ready policies Good, but longer to implement
Most Practical
ChatGPT o3 (Advanced reasoning)

Best balance of brevity + concrete governance (branch protection, CODEOWNERS, PR template, CI gates).

Most Thoughtful
ChatGPT Deep Research

Best when you need the “why,” alternatives, and (optionally) citations for training/governance docs.

Best For Quick Onboarding
ChatGPT 4.0 (Standard)

Fast, simple steps to get juniors productive without overwhelming them.

Reflection

  • How did the “intelligence level” affect quality? Higher levels added governance (branch protection, CI, CODEOWNERS), explicit trade-offs, and rationale — not just commands.
  • Which model for which challenge? Day-to-day policy/playbook → o3; Training and documentation with citations → Deep Research; Quick onboarding → 4.0.
  • What surprised you? Deep Research increased trust and teachability but slowed implementation; o3 required fewer prompts to produce copy-paste team rules.
Summary Insight Focused Scope
HSBR Note — Summary:

This study compared how three ChatGPT models approach the same complex workflow question. It revealed that ChatGPT o3 offered the most practical, ready-to-use team playbook, Deep Research provided the most insightful reasoning, and ChatGPT 4.0 Standard excelled at clear and simple onboarding.

Together, the results show how different intelligence levels shape both the depth and style of responses — from quick instruction to detailed strategic thinking. This page intentionally focuses only on Task 7 to keep the evaluation tight and comparable.