Google · LIVEBENCH 2026-06-25

Gemini 3.6 Flash High

Published benchmark results for this specific model configuration.

gemini-3.6-flash-high

Overall score73.6 / 100

Cost / successful task$0.235USD · benchmark workload

Weight accessNot reported

Where this configuration performs

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

Reasoning85.1

Coding77.9

Agentic Coding43.4

Mathematics86.4

Data Analysis63.0

Language83.9

Instruction Following75.4

Task-level results

theory of mind78.8zebra puzzle85.8spatial100.0logic with navigation76.0code generation77.5code completion78.3javascript63.6typescript26.7python40.0AMPS Hard98.0integrals with game63.0math comp94.1olympiad90.5consecutive events43.8tablejoin47.1tablereformat98.0connections100.0plot unscrambling63.7typos88.0paraphrase74.3simplify75.7story generation75.6summarize75.8

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