Leaderboards
One board per benchmark. Scores are never merged across benchmarks, because each one measures something different on a different scale.
| Benchmark | Use case | Models | Leader | Latest source |
|---|---|---|---|---|
| AA-AnalystAgent | Spreadsheets and accounting | 3 | Gemini 3.7 Flash (High) (60 pass^5 %) | date not stated |
| Advanced Financial Reasoning at Scale (CFA Level III evaluation) | Finance knowledge | 8 | o4-mini (79.1 overall % (mean of MCQ and essay)) | 22 Sept 2025 |
| Agent Market Arena (AMA) | Trading and portfolios | 10 | InvestorAgent / GPT-4.1 (40.8 cumulative return % (TSLA)) | 30 Oct 2025 |
| AI-Trader | Trading and portfolios | 6 | MiniMax-M2 (9.6 cumulative return %) | 1 Dec 2025 |
| APEX-Accounting | Spreadsheets and accounting | 10 | Gemini 4 Argon (71.5 Mean Score (% criteria passed)) | 1 Nov 2026 |
| BizFinBench | Finance knowledge | 10 | ChatGPT-o3 (73.9 average score across 9 tasks) | 26 May 2025 |
| BlueFin | Spreadsheets and accounting | 5 | GPT 5.5 (49.6 criteria-weighted score % (overall)) | May 2026 |
| CAIA (Crypto AI Agent Benchmark) | Live market data | 10 | gpt-5 (67.4 accuracy % (majority vote, 5 runs)) | 17 Jan 2026 |
| ClaimBench | Verifiability | 10 | d1-ultra (100 accuracy %) | 23 Oct 2026 |
| CorpFin v2 | Filings Q&A | 10 | Claude Opus 5 (73.2 accuracy %) | 12 Aug 2026 |
| Deep FinResearch Bench | Analyst research | 4 | OpenAI (86 overall factual claim rate %) | 22 Apr 2026 |
| Excel Modeling Benchmark (EMB) | Spreadsheets and accounting | 10 | Claude Fable 5.1 (76.7 accuracy %) | 29 Oct 2026 |
| FAITH | Verifiability | 10 | Claude Sonnet 4 (95.6 overall accuracy %) | 24 Oct 2025 |
| Finance Agent Benchmark | Analyst research | 10 | Claude Opus 5.5 (67.8 Partial Credit accuracy %) | 29 Oct 2026 |
| FinanceBench | Filings Q&A | 3 | GPT-4-Turbo (85 % correct (150-question sample)) | 20 Nov 2023 |
| FinanceQA | Filings Q&A | 4 | o1 (0.487 accuracy (total, 0-1)) | 30 Jan 2025 |
| Finch: Finance & Accounting Cowork Benchmark | Spreadsheets and accounting | 2 | GPT 5.1 Pro (38.4 workflow pass rate % (human eval)) | Apr 2026 |
| FinDeepResearch | Analyst research | 10 | FinAgentTest (41.9 overall accuracy score (0-100)) | 23 Mar 2026 |
| FinLFQA | Verifiability | 8 | GPT-4o (56.5 evidence attribution F1) | 7 Oct 2025 |
| FinResearchBench | Analyst research | 8 | Gemini (76.6 Final Score (0-100)) | 20 Oct 2025 |
| FinSafetyBench | Advice and safety | 6 | LLaMA-3 (32.3 attack success rate % (financial crimes, English)) | 1 May 2026 |
| FinSearchComp | Live market data | 10 | grok-4 (68.9 Global subset average accuracy %) | 16 Sept 2025 |
| FinTrust | Advice and safety | 11 | o4-mini (1 safety harmfulness score (1-10, GPT-4.1 mini judge)) | 17 Oct 2025 |
| FrontierFinance | Analyst research | 10 | Samaya Sterling (58.9 rubrics qualification rate %) | 28 Oct 2026 |
| InvestorBench | Trading and portfolios | 10 | Qwen2.5-72B-Instruct (46.2 cumulative return %) | 24 Dec 2024 |
| IPO Finance Agent | Filings Q&A | 6 | GLM-5.2 (79.8 mean rubric score %) | 30 Jun 2026 |
| K-FinHallu | Verifiability | 10 | Gemini 2.5 Flash (0.860 binary detection F1) | 28 May 2026 |
| KTD-Fin | Trading and portfolios | 10 | Qwen3.6-Plus (85.3 total return %) | 27 May 2026 |
| LiveTradeBench | Trading and portfolios | 10 | GPT-4.1 (6.3 cumulative return %) | 5 Nov 2025 |
| Market-Bench | Trading and portfolios | 10 | grok-4 (443.2 mean MAE) | date not stated |
| Open FinLLM Leaderboard | Finance knowledge | 10 | GPT-4o (46.0 balanced score (mean of text/vision/audio averages)) | 16 Apr 2026 |
| PolyBench | Trading and portfolios | 7 | MiMo-V2-Flash (17.6 Confidence-Weighted Return (CWR) %) | 3 Apr 2026 |
| PortBench | Trading and portfolios | 10 | GLM-5.1 (0.384 CEPS (0-1)) | 9 Oct 2026 |
| PRBench Finance | Advice and safety | 10 | Muse Spark 1.3 (59.5 weighted rubric score (0-100)) | date not stated |
| Reasoning Models Ace the CFA Exams (CFA mock-exam evaluation) | Finance knowledge | 6 | Gemini 3.0 Pro (97.6 Level I accuracy %) | 9 Dec 2025 |
| StockBench | Trading and portfolios | 11 | Qwen3-235B-Think (2.5 final return %) | 2 Mar 2026 |
| Tax Agent Bench | Finance knowledge | 10 | Qwen 3.8 Max 0902 (49.7 accuracy % (all-pass x citation factor)) | 29 Oct 2026 |
| Vals Finance Index | Analyst research | 10 | Gemini 4 Carbon (66.3 index accuracy %) | 29 Oct 2026 |