OpenAI · LIVEBENCH 2026-06-25

GPT-5.4 Nano xHigh

Published benchmark results for this specific model configuration.

gpt-5.4-nano-xhigh

Overall score69.6 / 100

Cost / successful task$0.091USD · benchmark workload

Weight accessNot reported

Where this configuration performs

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

Reasoning81.1

Coding70.8

Agentic Coding46.8

Mathematics91.0

Data Analysis67.6

Language62.5

Instruction Following67.2

Task-level results

theory of mind80.8zebra puzzle97.6spatial82.0logic with navigation64.0code generation74.3code completion67.4javascript63.6typescript26.7python50.0AMPS Hard98.0integrals with game82.8math comp95.1olympiad88.0consecutive events54.3tablejoin50.6tablereformat98.0connections94.4plot unscrambling39.1typos54.0paraphrase66.6simplify62.8story generation67.7summarize71.8

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