Meta · LIVEBENCH 2026-06-25

Muse Spark 1.3 xHigh Effort

Published benchmark results for this specific model configuration.

muse-spark-1.3-xhigh

Overall score81.6 / 100

Cost / successful task$0.219USD · benchmark workload

Weight accessNot reported

Where this configuration performs

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

Reasoning89.7

Coding81.1

Agentic Coding64.1

Mathematics95.9

Data Analysis79.6

Language82.8

Instruction Following78.0

Task-level results

theory of mind84.6zebra puzzle98.0spatial100.0logic with navigation76.0code generation81.7code completion80.4javascript77.3typescript60.0python55.0AMPS Hard99.0integrals with game96.0math comp97.1olympiad91.7consecutive events90.5tablejoin48.2tablereformat100.0connections94.0plot unscrambling68.4typos86.0paraphrase78.6simplify72.5story generation77.5summarize83.4

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