aitoolcompass.

Find your next AI tool

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

← All model benchmarks

Anthropic · LIVEBENCH 2026-11-06

claude-haiku-5.5

Published benchmark results for this specific model configuration.

claude-haiku-5.5
Overall score70.6 / 100
Cost / successful task$1.235USD · benchmark workload
Weight accessNot reported

Where this configuration performs

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

Long Agentic47.9
Reasoning82.4
Coding77.6
Browser69.5
Data Analysis53.1
Research84.3
Design79.3

Task-level results

long agentic47.9reasoning82.4coding77.6browser69.5data analysis53.1research84.3design79.3

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