# Finance benchmarks Find tests for financial research, document analysis, accounting, and tax work. Canonical page: https://benchlm.ai/benchmarks/collections/finance 11 records. These tests evaluate specific tasks and agent setups. They do not establish investment performance or a single best model for finance. Records keep metrics and harness variants separate. Source rows can include different model and agent configurations; linked profiles count distinct public model identities. Public model rows link to original published artifacts, which may include publisher comparisons of other providers' models. These counts do not represent independent evaluations. Dates describe the stored source record. ## Financial research and reasoning | Benchmark | Stored coverage | Task and run conditions | Original source | |---|---|---|---| | [Finance Agent v2](/benchmarks/financeagentv2) | 78 stored source rows; 65 linked model profiles | The default Vals tool harness, partial-credit score, and all-pass score stay distinct. Version 1.1 appears below as historical evidence. Financial analyst task categories Source update in snapshot: 2026-10-12; Source snapshot captured: 2026-10-15T18:47:28.403Z | [Original source](https://www.vals.ai/benchmarks/fabv2) | | [PRBench Finance](/benchmarks/scale-prbench-finance) | 37 stored source rows; 19 linked model profiles | Published professional-reasoning configurations remain separate from finance-agent runs. 37 published rows Snapshot label: October 15, 2026 snapshot | [Original source](https://labs.scale.com/leaderboard/prbench-finance) | | [FinanceArena](/benchmarks/financearena) | 19 stored source rows; 14 linked model profiles | The source table is labeled FinanceQA, Assumption-Based. It does not publish a task count or per-row run settings. Professional financial-analysis questions Snapshot label: October 12, 2026 snapshot | [Original source](https://www.afterquery.com/leaderboard/finance-arena) | ## Financial documents and modeling | Benchmark | Stored coverage | Task and run conditions | Original source | |---|---|---|---| | [CorpFin v2](/benchmarks/valscorpfinv2) | 134 stored source rows; 77 linked model profiles | Credit-agreement results preserve exact-page and context-budget views. Credit-agreement understanding tasks Source update in snapshot: 2026-08-12; Source snapshot captured: 2026-10-15T18:47:28.403Z | [Original source](https://www.vals.ai/benchmarks/corp_fin_v2) | | [MortgageTax](/benchmarks/valsmortgagetax) | 98 stored source rows; 58 linked model profiles | Semantic extraction and numerical extraction remain separate task views. Mortgage and tax extraction tasks Source update in snapshot: 2026-09-01; Source snapshot captured: 2026-10-15T18:47:28.403Z | [Original source](https://www.vals.ai/benchmarks/mortgage_tax) | | [EMB](/benchmarks/emb) | 74 stored source rows; 62 linked model profiles | Spreadsheet-modeling evidence comes from the published agent setup. Excel-based financial modeling tasks Source update in snapshot: 2026-10-12; Source snapshot captured: 2026-10-15T18:47:28.403Z | [Original source](https://www.vals.ai/benchmarks/emb) | ## Accounting, banking, and tax workflows | Benchmark | Stored coverage | Task and run conditions | Original source | |---|---|---|---| | [Ramp Accounting Bench](/benchmarks/ramp-accounting-bench) | 16 stored source rows; 16 linked model profiles | A fixed accounting-agent setup reports mean criterion-level scores across three attempts. 137 accounting tasks ยท 3 attempts per task Snapshot label: October 14, 2026 snapshot | [Original source](https://labs.ramp.com/ramp-accounting-bench#explore) | | [BankerToolBench](/benchmarks/bankertoolbench) | 2 public model rows linked to original artifacts | Sparse provider-reported banking-tool results are a starting point for inspection. Finance and banking tool-use tasks | [Original source](https://huggingface.co/MiniMaxAI/MiniMax-M3) | | [Tax Agent Bench](/benchmarks/valstaxagentbench) | 69 stored source rows; 60 linked model profiles | Overall and all-pass results use a dedicated US tax harness. US tax research, calculations, forms, filings, and precedent analysis Source update in snapshot: 2026-10-12; Source snapshot captured: 2026-10-15T18:47:28.403Z | [Original source](https://www.vals.ai/benchmarks/tax_agent_bench) | | [TaxEval v2](/benchmarks/taxevalv2) | 145 stored source rows; 82 linked model profiles | Read tax-task conditions alongside each result. Tax question answering and response evaluation Source update in snapshot: 2026-09-01; Source snapshot captured: 2026-10-15T18:47:28.403Z | [Original source](https://www.vals.ai/benchmarks/tax_eval_v2) | ## Historical evidence | Benchmark | Stored coverage | Task and run conditions | Original source | |---|---|---|---| | [Finance Agent v1.1](/benchmarks/valsfinanceagentv1) | 51 stored source rows; 31 linked model profiles | Archived source board. Archived Finance Agent v1.1 has a different task taxonomy and grading protocol from v2. 11 financial analyst task views Source update in snapshot: 2026-06-04; Source snapshot captured: 2026-10-15T18:47:28.403Z | [Original source](https://www.vals.ai/benchmarks/finance_agent) | ## Other collections - [Legal benchmarks](/benchmarks/collections/legal) - [Healthcare benchmarks](/benchmarks/collections/healthcare) [All benchmarks](/benchmarks)