xAI · LIVEBENCH 2026-06-25

Grok 4.5

Published benchmark results for this specific model configuration.

grok-4.5

Overall score75.8 / 100

Cost / successful task$0.131USD · benchmark workload

Weight accessNot reported

Where this configuration performs

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

Reasoning87.2

Coding68.6

Agentic Coding56.5

Mathematics90.8

Data Analysis73.0

Language82.8

Instruction Following71.5

Task-level results

theory of mind82.7zebra puzzle94.0spatial100.0logic with navigation72.0code generation67.6code completion69.6javascript72.7typescript36.7python60.0AMPS Hard99.0integrals with game78.0math comp97.1olympiad89.2consecutive events75.5tablejoin43.6tablereformat100.0connections98.0plot unscrambling66.4typos84.0paraphrase69.2simplify70.2story generation73.6summarize73.2

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