GPT-5.3-Codex

OpenAI llm

Advances GPT-5.2-Codex's agentic coding together with GPT-5.2's reasoning in one model; OpenAI states early versions helped debug its own training.

overall ⌄

—

nothing published

price · $/M tokens

1.75 / 14

in / out, as of 2026-07-25

to run it yourself ⌄

API only

no weights to run

Benchmarks

What its maker published, and what anyone else measured. Every figure links the document it came from.

benchmark	published	measured
GPQA	—	91.5
HLE	—	39.9

Nothing is both cheaper and ranked as high. This model is on the value frontier — whatever you pay for it, you are paying for something.

↑ merge GPT-5.2