FinBalance: leaderboard
FinBalance (2026) is a multi-document accounting reconciliation benchmark: from a bundle of source documents with distractors, a model posts journal entries citing their documents, aggregates them into a balance sheet and flags inconsistencies, scored against a deterministic ledger. This board is exact balance-sheet accuracy in percent on the 480 standard records: the reported balance sheet must match the ground truth exactly; higher is better.
Metric: Exact balance-sheet accuracy (%): share of records whose reported final balance sheet matches the ledger ground truth exactly (BS exact); 480 standard records of the 710-record core evaluation split (source-document bundles with distractors across eight industries, three period types and five difficulty levels), baseline prompt, temperature 0, one pass; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 6 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5 (Low) | 46.5 | #167 (GPT-5) |
| 2 | Grok 4.3 (Non-reasoning) | 31 | #333 (Grok 4.3) |
| 3 | Gemini 3 Flash (Minimal) | 23.5 | #164 (Gemini 3 Flash) |
| 4 | Claude Haiku 4.5 | 2.9 | #568 |
| 5 | Qwen 3 235B A22B 2507 Instruct | 1.3 | #619 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=finbalance · How It Works · Data refreshed daily, snapshot 2026-11-13.