← Latest products

Do Codex skills save tokens? A six-run task-size benchmark

Explore six controlled GPT-5.6-sol runs showing when an engineering skill helped—and when it added overhead.

Visit codex-howto-benchmark.nguyenvantamdk2.chatgpt.siteexternal site

Reactions (1)

Similar products

by description similarity

  1. OpenSkills

    news.ycombinator.com

    Stop bloating your LLM context with unused instructions

    Hacker News9 pts

  2. Typeui.sh

    typeui.sh

    generate skill files for design system specifications

    Hacker News10 pts

  3. Ingot

    github.com

    evidence-gated optimization and version control for agent skills

    Hacker News7 pts

  4. Predict GPT-5 skills with a community AI benchmark

    news.ycombinator.com

    I’m Andrew, co-founder of Recall.

    Hacker News13 pts

  5. Simpler

    simplerlist.com

    Your GPT-3 Task Planner

    Hacker News7 pts