OpenAI · LIVEBENCH 2026-06-25

GPT-5.6 Sol Max Effort

Published benchmark results for this specific model configuration.

gpt-5.6-sol-max

Overall score81.1 / 100

Cost / successful task$0.515USD · benchmark workload

Weight accessNot reported

Where this configuration performs

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

Reasoning91.7

Coding83.9

Agentic Coding56.2

Mathematics96.2

Data Analysis79.8

Language87.7

Instruction Following71.8

Task-level results

theory of mind84.6zebra puzzle100.0spatial100.0logic with navigation82.0code generation83.1code completion84.8javascript63.6typescript50.0python55.0AMPS Hard98.0integrals with game100.0math comp95.1olympiad91.7consecutive events90.2tablejoin49.3tablereformat100.0connections100.0plot unscrambling79.1typos84.0paraphrase71.5simplify70.5story generation75.9summarize69.4

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