DualEntry Accounting AI Benchmark
Finance & Accounting
DualEntry's DualBench100: 101 synthetic accounting cases across eight areas (transaction classification, journal entries, accounts payable and receivable, bank reconciliation, financial reporting, month-end close and accounting knowledge), each run three times in a freshly restored test company with the same accounting tools. Accuracy is the share of cases passed, by deterministic checks against fixed reference answers, in at least two of three runs.
Top models (higher is better)
| Model | Score |
|---|---|
| Gemini 4 Argon | 91.1 |
| GLM-5.3 Flash | 88.1 |
| Grok 4.7 | 88.1 |
| Kimi K3.1 Pro | 86.1 |
| Claude Fable 5.1 | 85.1 |
| DeepSeek V4.1 Flash | 84.2 |
| Kimi K3 | 84.2 |
| Gemini 3.8 Flash | 83.2 |
| Grok 4.5 | 83.2 |
| Muse Wonder | 83.2 |
| Claude Opus 5.5 | 82.2 |
| DeepSeek V4.1 Flash 1017 | 82.2 |
| GLM-5.3 | 82.2 |
| GPT-6.1 Sol | 81.2 |
| GPT-6 Astra | 81.2 |
| Claude Sonnet 5.5 | 80.2 |
| Mistral Large 4 Preview | 80.2 |
| GPT-6.1 Terra | 78.2 |
| Mercury 3 | 78.2 |
| Claude Haiku 5.5 | 76.2 |
| GPT-6 Sol | 76.2 |
| GPT-6 Luna | 66.3 |