OpenAI · LIVEBENCH 2026-06-25

GPT-5.6 Luna Max Effort

Published benchmark results for this specific model configuration.

gpt-5.6-luna-max

Overall score73.6 / 100

Cost / successful task$0.169USD · benchmark workload

Weight accessNot reported

Where this configuration performs

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

Reasoning85.6

Coding82.9

Agentic Coding48.4

Mathematics87.2

Data Analysis78.0

Language72.6

Instruction Following60.1

Task-level results

theory of mind73.1zebra puzzle93.5spatial96.0logic with navigation80.0code generation78.9code completion87.0javascript63.6typescript36.7python45.0AMPS Hard98.0integrals with game70.0math comp92.2olympiad88.6consecutive events86.5tablejoin47.6tablereformat100.0connections96.5plot unscrambling51.2typos70.0paraphrase47.8simplify58.7story generation69.1summarize64.9

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