GPT-5.3-Codex: terminal-Bench 2.0 corrected from 78.4% to OpenAI’s published 77.3%
Tracker verifies GPT-5.3-Codex’s Terminal-Bench 2.0 score against OpenAI documentation
Terminal-Bench 2.0 corrected from 78.4% to OpenAI’s published 77.3% — verified by the OnAirToday AI Progress Tracker against OpenAI's published documentation.
OnAirToday
Newsroom
The OnAirToday AI Progress Tracker recorded a verified change to OpenAI's GPT-5.3-Codex (version gpt-5.3-codex): Terminal-Bench 2.0 corrected from 78.4% to OpenAI’s published 77.3%.
GPT-5.3-Codex is OpenAI’s coding-focused agentic model for software engineering and computer-based work. OpenAI reports faster Codex execution, stronger coding and reasoning capabilities, and improved support for steering ongoing tasks, while human review and restricted tool permissions remain important.
What changed in this review:
- The previously listed SWE-bench Verified figure is omitted because it lacks a model-specific published citation.
- Context-window and API-pricing figures are left unspecified without supporting model-specific documentation.
- Flaws and limitations updated
Teams building on GPT-5.3-Codex should confirm limits, pricing and benchmark coverage against the vendor's own documentation before relying on these figures.
The full profile — capabilities, benchmark citations, known flaws and risk guidance — is on the tracker at /ai-models/gpt-5-3-codex.
References:
- https://openai.com/index/introducing-gpt-5-3-codex/
- https://www.tbench.ai/leaderboard/terminal-bench/2.0
Generated insight by OnAirToday AI
The correction aligns the reported Terminal-Bench 2.0 score with OpenAI’s published documentation. Accurate benchmark figures matter when readers compare AI model performance.

Comments
Keep comments relevant and respectful. No links, advertising, self-promotion, or spam.