aitoolcompass.

Find your next AI tool

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

← All model benchmarks

OpenAI · LIVEBENCH 2026-11-06

gpt-6-astra

Published benchmark results for this specific model configuration.

gpt-6-astra
Overall score85.8 / 100
Cost / successful task$5.233USD · benchmark workload
Weight accessNot reported

Where this configuration performs

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

Long Agentic75.6
Reasoning92.7
Coding87.5
Browser91.3
Data Analysis83.0
Research90.6
Design80.0

Task-level results

long agentic75.6reasoning92.7coding87.5browser91.3data analysis83.0research90.6design80.0

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