OpenAI launches GPT-5-codex with a success rate of 74.5% in real world coding




  • GPT-5-Codex promises greater performance and successful rates
  • It is included for Plus, Pro, Business, Edu and Enterprise users
  • The model can use 93.7% less tokens in light tasks

Operai has shared more details about GPT-5-Codex, an specially designed version of GPT-5 specifically optimized for the coding of agent and software engineering of the real world, and we are in a pleasure when it comes to reliability and performance.

The Chatgpt manufacturer affirmed a Swe-Bench verified reference rate of 74.5%, with a refactorization performance improving 51.3% (compared to 33.9% in GPT-5).

Leave a Comment

Your email address will not be published. Required fields are marked *