AccountingBench — Analytics

Analytics Dashboard

—

Top Overall Score

—

Avg. Score (all models)

—

Tasks in Benchmark

—

Models Evaluated

—

Question Distribution

—

Tasks

API Cost Analysis

Prices as of March 2026 · sorted by cost/task

Model	In $/1M	Out $/1M	Cost/Task

Note: Cost per task for Kimi K2.6 could not be estimated due to technical issues.

Avg. Token Usage per Task

Total input + output tokens · 1 trial per task

Note: Avg. token usage per task for Kimi K2.6 could not be estimated due to technical issues.

Cost-Efficiency: Score vs. Avg. Cost per Task

Score vs. Output Speed

AccountingBench score vs. output tokens per second · Source: Artificial Analysis (March 2026)

Confidence–Performance Calibration

AccountingBench score versus model confidence · bins with fewer than three observations are omitted.

—

Category & Task Type

Category	Tasks	Share

Task Type	Tasks	Share

Question Format & Education Level

Question Format	Tasks	Share

Education Level	Tasks	Share

Regulatory Framework Distribution

Regulatory Framework	Tasks	Share