aitoolcompass.

Find your next AI tool

Search by product name or task. Press Escape to close.

← All model benchmarks

Unspecified · LIVEBENCH 2026-11-06

step-5-preview

Published benchmark results for this specific model configuration.

step-5-preview
Overall score58.2 / 100
Cost / successful task$1.280USD · benchmark workload
Weight accessNot reported

Where this configuration performs

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

Long Agentic19.4
Reasoning83.8
Coding69.4
Browser24.8
Data Analysis63.5
Research75.5
Design70.8

Task-level results

long agentic19.4reasoning83.8coding69.4browser24.8data analysis63.5research75.5design70.8

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