OpenAI · LIVEBENCH 2026-06-25

GPT-5.4 Mini xHigh

Published benchmark results for this specific model configuration.

gpt-5.4-mini-xhigh

Overall score66.4 / 100

Cost / successful task$0.334USD · benchmark workload

Weight accessNot reported

Where this configuration performs

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

Reasoning71.3

Coding71.6

Agentic Coding41.7

Mathematics78.5

Data Analysis70.8

Language71.0

Instruction Following59.8

Task-level results

theory of mind76.9zebra puzzle60.4spatial88.0logic with navigation60.0code generation67.7code completion75.6javascript50.0typescript20.0python55.0AMPS Hard97.0integrals with game37.0math comp94.1olympiad85.7consecutive events63.0tablejoin49.4tablereformat100.0connections99.0plot unscrambling43.9typos70.0paraphrase58.9simplify54.1story generation66.6summarize59.6

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