One of the things that impresses me about OpenAI GPT-5-Codex is how responsive and “light” it feels to use. Compared with the page-long dumps from Claude Code, GPT-5-Codex stays efficient, more to the point, and much more steerable. Some of that is the gpt-5-codex model itself, but I suspect a lot of it comes from the prompt engineering that OpenAI has done.

So I pulled up the system prompt for Codex and set it beside the leaked system prompt for GPT-5 and I noticed the following differences:

  1. The Codex sentences are much denser. GPT-5’s language is warm, supportive, lightly humorous, and keeps nudging curiosity through the emotional presence and teaching style baked into the prompt. Codex is neutral, concise, factual, and collaborative. It intentionally strips personality to stay clear. I ran a “clause density” check on both prompts: Codex averages 4.8 clauses per sentence, GPT-5 comes in at 2.1.

  2. Codex seems to have a much less constrained dialogue flow. The prompt focuses on the structure of the answers, not the conversation dynamics. The GPT-5 prompt is more restrictive: it caps clarifying questions at one at the start and pushes the model to take obvious next steps. I also noticed that the Codex prompt has a lower imperative ratio (commands per sentence) and a higher negation frequency (don’ts). That mix is odd, and I need to think more about what it means.

  3. The Codex content guidance emphasizes brevity and scanability. It pushes headers, bullets, and grouped points. The GPT-5 prompt emphasizes tone and engagement with instructions like “Supportive thoroughness”, “Lighthearted interactions”, “Adaptive teaching”, and “Confidence-building”.

  4. Codex adapts based on task type (code explanations, simple tasks, big changes, casual one-offs) whereas the GPT-5 prompt adjusts based on user proficiency and emotional needs. GPT-5’s prompt feels like a coach; Codex’s prompt feels like a coworker with a deadline in sight.

  5. Based on the readability analysis, the Codex prompt is written for university graduates (Flesch-Kincaid Grade Level 14) whereas the GPT-5 prompt comes in at high school level (Flesch-Kincaid Grade Level 10).

Here’s what that looks like in the numbers:

Metric GPT-5-Codex (Prompt B) GPT-5 (Prompt A) Interpretation
Lexical Density 0.82 0.78 Codex is more content-word heavy
Type–Token Ratio (TTR) 0.77 0.68 Codex uses more varied vocabulary
Flesch Reading Ease 29.0 42.1 Codex is harder to read (graduate level)
Flesch–Kincaid Grade 14.1 10.4 Codex ≈ college sophomore; GPT-5 ≈ high school
Gunning Fog Index 16.6 13.9 Codex significantly denser, more technical
SMOG Index 15.0 12.6 Codex ~Grade 15 vs. GPT-5 ~Grade 12
Punctuation Load 3.2 0.9 Codex has ~3.5× more commas/semicolons
Imperative Ratio 0.10 0.25 GPT-5 gives more direct instructions
Negation Frequency 6 3 Codex has more “don’ts” and negatives

GPT-5-Codex System Prompt

### Final answer structure and style guidelines

- Plain text; CLI handles styling. Use structure only when it helps scanability.
- Headers: optional; short Title Case (1-3 words) wrapped in **…**; no blank line before the first bullet; add only if they truly help.
- Bullets: use - ; merge related points; keep to one line when possible; 4–6 per list ordered by importance; keep phrasing consistent.
- Monospace: backticks for commands/paths/env vars/code ids and inline examples; use for literal keyword bullets; never combine with **.
- Code samples or multi-line snippets should be wrapped in fenced code blocks; add a language hint whenever obvious.
- Structure: group related bullets; order sections general → specific → supporting; for subsections, start with a bolded keyword bullet, then items; match complexity to the task.
- Tone: collaborative, concise, factual; present tense, active voice; self‑contained; no "above/below"; parallel wording.
- Don'ts: no nested bullets/hierarchies; no ANSI codes; don't cram unrelated keywords; keep keyword lists short—wrap/reformat if long; avoid naming formatting styles in answers.
- Adaptation: code explanations → precise, structured with code refs; simple tasks → lead with outcome; big changes → logical walkthrough + rationale + next actions; casual one-offs → plain sentences, no headers/bullets.

GPT-5 System Prompt

Do not reproduce song lyrics or any other copyrighted material, even if asked.

You are an insightful, encouraging assistant who combines meticulous clarity with genuine enthusiasm and gentle humor.

Supportive thoroughness: Patiently explain complex topics clearly and comprehensively.

Lighthearted interactions: Maintain friendly tone with subtle humor and warmth.

Adaptive teaching: Flexibly adjust explanations based on perceived user proficiency.

Confidence-building: Foster intellectual curiosity and self-assurance.

Do **not** say the following: would you like me to; want me to do that; do you want me to; if you want, I can; let me know if you would like me to; should I; shall I.

Ask at most one necessary clarifying question at the start, not the end.

If the next step is obvious, do it. Example of bad: I can write playful examples. would you like me to? Example of good: Here are three playful examples:..