GPT-5.4-Mini
The price-performance workhorse — 74.3% on DualEntry at a fraction of flagship cost.
Accounting overall
74.3%
Input / Output
$0.50 / $2.00 per MTok
Context
1M
Speed
~150 tok/s
Released
2026-03
Cutoff
2025-12
Eight accounting-task categories borrowed from DualEntry's 101-task benchmark. Measured where published, synthesized from adjacent benchmarks otherwise.
GPT-5.4-Mini is the clearest price-performance winner on the DualEntry leaderboard at 74.3% — within ~5 points of GPT-5.4 proper and within ~5 points of Opus 4.7, at a fraction of the cost. For production agentic accounting where every dollar of token cost is multiplied across thousands of monthly transactions, Mini is increasingly the default choice when absolute accuracy isn't mission-critical.
Where Mini underperforms: complex month-end close reasoning and multi-step financial reporting — consistent with the pattern across all current models, where the gap between frontier and smaller models widens on workflows that require sustained multi-step judgment.
Citations
- DualEntry benchmark (GPT-5.4-Mini 74.3%)dualentry.com/blog/claude-opus-4-7-accounting-ai-benchmark-results