Skip to content

GPT-5.3-Codex

Released 2026-02 · reasoning · 400K tokens · 3 benchmarks

Editorial notes

Lanzado 5 febrero 2026. Coding-specialist agentic 25% mas rapido que GPT-5.2-Codex. Combina coding frontier con reasoning de 5.2. Primer modelo OpenAI clasificado "High" en Cybersecurity (Preparedness Framework). GA en GitHub Copilot 9 feb 2026. Pricing $1.75/$14 per MTok.

Spec sheet

Company
OpenAI
Country
US
Type
reasoning
Release
2026-02
Context
400K tokens
Slug
gpt-5-3-codex

Benchmarks (3)

Cite this model
BibTeX · APA

BibTeX

@misc{frontier-gpt-5-3-codex,
  title  = {GPT-5.3-Codex},
  author = {{OpenAI}},
  year   = {2026},
  note   = {Frontier Benchmarks AI atlas. Accessed 2026-05-08},
  url    = {https://frontierbenchmarks.com/models/gpt-5-3-codex}
}

APA

OpenAI (2026). GPT-5.3-Codex [Large language model]. Frontier Benchmarks AI. Retrieved 2026-05-08, from https://frontierbenchmarks.com/models/gpt-5-3-codex

Citation reflects the atlas page, not the original model paper. For the paper, see the "Resources" section above.