OpenAI · LIVEBENCH 2026-06-25

GPT-5.2 Codex

Published benchmark results for this specific model configuration.

gpt-5.2-codex

Overall score74.0 / 100

Cost / successful task$0.187USD · benchmark workload

Weight accessNot reported

Where this configuration performs

Category averages from the same release, on a 0–100 scale.

Reasoning77.7

Coding83.6

Agentic Coding49.4

Mathematics88.8

Data Analysis78.2

Language73.7

Instruction Following66.4

Task-level results

theory of mind78.8zebra puzzle70.0spatial94.0logic with navigation68.0code generation80.3code completion87.0javascript68.2typescript20.0python60.0AMPS Hard98.0integrals with game75.0math comp95.1olympiad87.0consecutive events89.0tablejoin45.6tablereformat100.0connections95.0plot unscrambling56.0typos70.0paraphrase67.0simplify60.6story generation74.8summarize63.4

Source files: Scores ↗ Categories ↗ Costs ↗ . Missing cost is unknown, never free. Compare configurations within the same release.