Unspecified · LIVEBENCH 2026-06-25

mistral-large-4-high

Published benchmark results for this specific model configuration.

mistral-large-4-high

Overall score71.8 / 100

Cost / successful task$0.139USD · benchmark workload

Weight accessNot reported

Where this configuration performs

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

Reasoning83.9

Coding77.2

Agentic Coding57.2

Mathematics93.6

Data Analysis76.5

Language49.6

Instruction Following64.8

Task-level results

theory of mind80.8zebra puzzle93.0spatial94.0logic with navigation68.0code generation76.1code completion78.3javascript68.2typescript43.3python60.0AMPS Hard98.0integrals with game90.0math comp97.1olympiad89.3consecutive events80.5tablejoin49.1tablereformat100.0connections48.0plot unscrambling40.9typos60.0paraphrase64.0simplify66.2story generation67.5summarize61.6

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