# Best AI for Finance knowledge (November 2026)

The leader on each benchmark in this use case, never combined. Data refreshed 2 Nov 2026.

## Picks

- BizFinBench: ChatGPT-o3: 73.9 average score across 9 tasks (Paper Table 3, 26 May 2025), https://benchmarks.finance/leaderboards/bizfinbench/
- Open FinLLM Leaderboard: GPT-4o: 46.0 balanced score (mean of text/vision/audio averages) (Open FinLLM Leaderboard (scores.json), 16 Apr 2026), https://benchmarks.finance/leaderboards/open-finllm-leaderboard/
- Tax Agent Bench: Qwen 3.8 Max 0902: 49.7 accuracy % (all-pass x citation factor) (Vals.ai leaderboard, 29 Oct 2026), https://benchmarks.finance/leaderboards/vals-tax-agent-bench/
- Advanced Financial Reasoning at Scale (CFA Level III evaluation): o4-mini: 79.1 overall % (mean of MCQ and essay) (Paper Table 1, 22 Sept 2025), https://benchmarks.finance/leaderboards/cfa-level-3-at-scale/
- Reasoning Models Ace the CFA Exams (CFA mock-exam evaluation): Gemini 3.0 Pro: 97.6 Level I accuracy % (Paper Table 5, 9 Dec 2025), https://benchmarks.finance/leaderboards/cfa-reasoning-models-ace/
- TaxEval v2: Muse Spark 1.2: 80.4 accuracy % (v2) (Vals.ai TaxEval v2 leaderboard, 1 Sept 2026), https://benchmarks.finance/leaderboards/vals-tax-eval-v2/
- τ³-Banking: Qwen 3.8 Max: 55.2 pass^1 % (τ-bench leaderboard (Sierra), 4 Aug 2026), https://benchmarks.finance/leaderboards/tau3-banking/

## BizFinBench: average score across 9 tasks (higher is better, 10 models)

| Rank | Model | Provider | Score | Metric | Setting | Reported by | Source | URL | Date |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| 1 | ChatGPT-o3 | OpenAI | 73.9 | average score across 9 tasks |  | Reported by the authors | Paper Table 3 | https://arxiv.org/html/2505.19457v1 | 26 May 2025 (checked 1 Nov 2026) |
| 2 | DeepSeek-R1 (671B) | DeepSeek | 73.0 | average score across 9 tasks |  | Reported by the authors | Paper Table 3 | https://arxiv.org/html/2505.19457v1 | 26 May 2025 (checked 1 Nov 2026) |
| 3 | GPT-4o | OpenAI | 71.8 | average score across 9 tasks |  | Reported by the authors | Paper Table 3 | https://arxiv.org/html/2505.19457v1 | 26 May 2025 (checked 1 Nov 2026) |
| 4 | DeepSeek-V3 (671B) | DeepSeek | 71.6 | average score across 9 tasks |  | Reported by the authors | Paper Table 3 | https://arxiv.org/html/2505.19457v1 | 26 May 2025 (checked 1 Nov 2026) |
| 5 | ChatGPT-o4-mini | OpenAI | 71.3 | average score across 9 tasks |  | Reported by the authors | Paper Table 3 | https://arxiv.org/html/2505.19457v1 | 26 May 2025 (checked 1 Nov 2026) |
| 6 | Gemini-2.0-Flash | Google | 69.8 | average score across 9 tasks |  | Reported by the authors | Paper Table 3 | https://arxiv.org/html/2505.19457v1 | 26 May 2025 (checked 1 Nov 2026) |
| 7 | Qwen2.5-VL-32B | Qwen | 68.6 | average score across 9 tasks |  | Reported by the authors | Paper Table 3 | https://arxiv.org/html/2505.19457v1 | 26 May 2025 (checked 1 Nov 2026) |
| 8 | Qwen3-32B | Qwen | 68.3 | average score across 9 tasks |  | Reported by the authors | Paper Table 3 | https://arxiv.org/html/2505.19457v1 | 26 May 2025 (checked 1 Nov 2026) |
| 9 | Qwen2.5-72B-Instruct | Qwen | 67.7 | average score across 9 tasks |  | Reported by the authors | Paper Table 3 | https://arxiv.org/html/2505.19457v1 | 26 May 2025 (checked 1 Nov 2026) |
| 10 | Qwen2.5-VL-72B | Qwen | 67.5 | average score across 9 tasks |  | Reported by the authors | Paper Table 3 | https://arxiv.org/html/2505.19457v1 | 26 May 2025 (checked 1 Nov 2026) |

## Open FinLLM Leaderboard: balanced score (mean of text/vision/audio averages) (higher is better, 10 models)

| Rank | Model | Provider | Score | Metric | Setting | Reported by | Source | URL | Date |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| 1 | GPT-4o | OpenAI | 46.0 | balanced score (mean of text/vision/audio averages) | MultiFinBen Table 3 data | Official board | Open FinLLM Leaderboard (scores.json) | https://huggingface.co/spaces/TheFinAI/Open-FinLLM-Leaderboard | 16 Apr 2026 (checked 1 Nov 2026) |
| 2 | Qwen2.5-Omni-7B | Qwen | 33.5 | balanced score (mean of text/vision/audio averages) | MultiFinBen Table 3 data | Official board | Open FinLLM Leaderboard (scores.json) | https://huggingface.co/spaces/TheFinAI/Open-FinLLM-Leaderboard | 16 Apr 2026 (checked 1 Nov 2026) |
| 3 | Llama-4-Scout-17B-16E | Meta | 20.9 | balanced score (mean of text/vision/audio averages) | MultiFinBen Table 3 data | Official board | Open FinLLM Leaderboard (scores.json) | https://huggingface.co/spaces/TheFinAI/Open-FinLLM-Leaderboard | 16 Apr 2026 (checked 1 Nov 2026) |
| 4 | Whisper-V3 | OpenAI | 17.2 | balanced score (mean of text/vision/audio averages) | MultiFinBen Table 3 data | Official board | Open FinLLM Leaderboard (scores.json) | https://huggingface.co/spaces/TheFinAI/Open-FinLLM-Leaderboard | 16 Apr 2026 (checked 1 Nov 2026) |
| 5 | Qwen2-Audio-7B-Instruct | Qwen | 16.7 | balanced score (mean of text/vision/audio averages) | MultiFinBen Table 3 data | Official board | Open FinLLM Leaderboard (scores.json) | https://huggingface.co/spaces/TheFinAI/Open-FinLLM-Leaderboard | 16 Apr 2026 (checked 1 Nov 2026) |
| 6 | Qwen2-Audio-7B | Qwen | 16.0 | balanced score (mean of text/vision/audio averages) | MultiFinBen Table 3 data | Official board | Open FinLLM Leaderboard (scores.json) | https://huggingface.co/spaces/TheFinAI/Open-FinLLM-Leaderboard | 16 Apr 2026 (checked 1 Nov 2026) |
| 7 | Gemma-3-27B | Google | 15.7 | balanced score (mean of text/vision/audio averages) | MultiFinBen Table 3 data | Official board | Open FinLLM Leaderboard (scores.json) | https://huggingface.co/spaces/TheFinAI/Open-FinLLM-Leaderboard | 16 Apr 2026 (checked 1 Nov 2026) |
| 8 | Llama-3.1-70B | Meta | 14.1 | balanced score (mean of text/vision/audio averages) | MultiFinBen Table 3 data | Official board | Open FinLLM Leaderboard (scores.json) | https://huggingface.co/spaces/TheFinAI/Open-FinLLM-Leaderboard | 16 Apr 2026 (checked 1 Nov 2026) |
| 9 | Gemma-3-4B | Google | 12.7 | balanced score (mean of text/vision/audio averages) | MultiFinBen Table 3 data | Official board | Open FinLLM Leaderboard (scores.json) | https://huggingface.co/spaces/TheFinAI/Open-FinLLM-Leaderboard | 16 Apr 2026 (checked 1 Nov 2026) |
| 10 | Qwen2.5-32B | Qwen | 12.7 | balanced score (mean of text/vision/audio averages) | MultiFinBen Table 3 data | Official board | Open FinLLM Leaderboard (scores.json) | https://huggingface.co/spaces/TheFinAI/Open-FinLLM-Leaderboard | 16 Apr 2026 (checked 1 Nov 2026) |

## Tax Agent Bench: accuracy % (all-pass x citation factor) (higher is better, 10 models)

| Rank | Model | Provider | Score | Metric | Setting | Reported by | Source | URL | Date |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| 1 | Qwen 3.8 Max 0902 | Qwen | 49.7 | accuracy % (all-pass x citation factor) | test set | Official board | Vals.ai leaderboard | https://www.vals.ai/benchmarks/tax_agent_bench | 29 Oct 2026 (checked 1 Nov 2026) |
| 2 | Claude Fable 5.1 | Anthropic | 49.2 | accuracy % (all-pass x citation factor) | test set | Official board | Vals.ai leaderboard | https://www.vals.ai/benchmarks/tax_agent_bench | 29 Oct 2026 (checked 1 Nov 2026) |
| 3 | Gemini 4 Carbon | Google | 48.9 | accuracy % (all-pass x citation factor) | test set | Official board | Vals.ai leaderboard | https://www.vals.ai/benchmarks/tax_agent_bench | 29 Oct 2026 (checked 1 Nov 2026) |
| 4 | Claude Opus 5.5 | Anthropic | 45.6 | accuracy % (all-pass x citation factor) | test set | Official board | Vals.ai leaderboard | https://www.vals.ai/benchmarks/tax_agent_bench | 29 Oct 2026 (checked 1 Nov 2026) |
| 5 | Claude Opus 5 | Anthropic | 45.5 | accuracy % (all-pass x citation factor) | test set | Official board | Vals.ai leaderboard | https://www.vals.ai/benchmarks/tax_agent_bench | 29 Oct 2026 (checked 1 Nov 2026) |
| 6 | Gemini 4 Argon | Google | 44.6 | accuracy % (all-pass x citation factor) | test set | Official board | Vals.ai leaderboard | https://www.vals.ai/benchmarks/tax_agent_bench | 29 Oct 2026 (checked 1 Nov 2026) |
| 7 | DeepSeek V4.1 Flash 1017 | DeepSeek | 44.4 | accuracy % (all-pass x citation factor) | test set | Official board | Vals.ai leaderboard | https://www.vals.ai/benchmarks/tax_agent_bench | 29 Oct 2026 (checked 1 Nov 2026) |
| 8 | Wonder Full | Not stated | 43 | accuracy % (all-pass x citation factor) | test set | Official board | Vals.ai leaderboard | https://www.vals.ai/benchmarks/tax_agent_bench | 29 Oct 2026 (checked 1 Nov 2026) |
| 9 | Muse Spark 1.3 Max | Meta | 42.4 | accuracy % (all-pass x citation factor) | test set | Official board | Vals.ai leaderboard | https://www.vals.ai/benchmarks/tax_agent_bench | 29 Oct 2026 (checked 1 Nov 2026) |
| 10 | Claude Sonnet 5.5 | Anthropic | 42.0 | accuracy % (all-pass x citation factor) | test set | Official board | Vals.ai leaderboard | https://www.vals.ai/benchmarks/tax_agent_bench | 29 Oct 2026 (checked 1 Nov 2026) |

## Advanced Financial Reasoning at Scale (CFA Level III evaluation): overall % (mean of MCQ and essay) (higher is better, 8 models)

| Rank | Model | Provider | Score | Metric | Setting | Reported by | Source | URL | Date |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| 1 | o4-mini | OpenAI | 79.1 | overall % (mean of MCQ and essay) | MCQ Self-Discover; essay CoT-SC N=3 | Reported by the authors | Paper Table 1 | https://arxiv.org/html/2507.02954v2 | 22 Sept 2025 (checked 1 Nov 2026) |
| 2 | Gemini 2.5 Flash | Google | 77.3 | overall % (mean of MCQ and essay) | MCQ zero-shot; essay CoT-SC N=3 | Reported by the authors | Paper Table 1 | https://arxiv.org/html/2507.02954v2 | 22 Sept 2025 (checked 1 Nov 2026) |
| 3 | Gemini 2.5 Pro | Google | 75.9 | overall % (mean of MCQ and essay) | MCQ zero-shot; essay CoT-SC N=3 | Reported by the authors | Paper Table 1 | https://arxiv.org/html/2507.02954v2 | 22 Sept 2025 (checked 1 Nov 2026) |
| 4 | Claude Opus 4 | Anthropic | 74.9 | overall % (mean of MCQ and essay) | MCQ CoT-SC N=3; essay CoT-SC N=3 | Reported by the authors | Paper Table 1 | https://arxiv.org/html/2507.02954v2 | 22 Sept 2025 (checked 1 Nov 2026) |
| 5 | o3-mini | OpenAI | 74.4 | overall % (mean of MCQ and essay) | MCQ CoT-SC N=5; essay Self-Discover or CoT-SC N=5 | Reported by the authors | Paper Table 1 | https://arxiv.org/html/2507.02954v2 | 22 Sept 2025 (checked 1 Nov 2026) |
| 6 | Grok-3 | xAI | 73.8 | overall % (mean of MCQ and essay) | MCQ CoT-SC N=3; essay CoT-SC N=5 | Reported by the authors | Paper Table 1 | https://arxiv.org/html/2507.02954v2 | 22 Sept 2025 (checked 1 Nov 2026) |
| 7 | Claude Sonnet 4 | Anthropic | 73.4 | overall % (mean of MCQ and essay) | MCQ CoT-SC N=3; essay CoT-SC N=3 | Reported by the authors | Paper Table 1 | https://arxiv.org/html/2507.02954v2 | 22 Sept 2025 (checked 1 Nov 2026) |
| 8 | Grok-3 Mini Low Effort | xAI | 72.4 | overall % (mean of MCQ and essay) | MCQ CoT-SC N=3; essay CoT-SC N=3 | Reported by the authors | Paper Table 1 | https://arxiv.org/html/2507.02954v2 | 22 Sept 2025 (checked 1 Nov 2026) |

## Reasoning Models Ace the CFA Exams (CFA mock-exam evaluation): Level I accuracy % (higher is better, 6 models)

| Rank | Model | Provider | Score | Metric | Setting | Reported by | Source | URL | Date |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| 1 | Gemini 3.0 Pro | Google | 97.6 | Level I accuracy % | zero-shot | Reported by the authors | Paper Table 5 | https://arxiv.org/html/2512.08270v1 | 9 Dec 2025 (checked 1 Nov 2026) |
| 2 | GPT-5 | OpenAI | 96.1 | Level I accuracy % | zero-shot | Reported by the authors | Paper Table 5 | https://arxiv.org/html/2512.08270v1 | 9 Dec 2025 (checked 1 Nov 2026) |
| 3 | Gemini 2.5 Pro | Google | 95.7 | Level I accuracy % | zero-shot | Reported by the authors | Paper Table 5 | https://arxiv.org/html/2512.08270v1 | 9 Dec 2025 (checked 1 Nov 2026) |
| 4 | Grok 4 | xAI | 94.8 | Level I accuracy % | zero-shot | Reported by the authors | Paper Table 5 | https://arxiv.org/html/2512.08270v1 | 9 Dec 2025 (checked 1 Nov 2026) |
| 5 | Claude Opus 4.1 | Anthropic | 94.6 | Level I accuracy % | zero-shot | Reported by the authors | Paper Table 5 | https://arxiv.org/html/2512.08270v1 | 9 Dec 2025 (checked 1 Nov 2026) |
| 6 | DeepSeek-V3.1 | DeepSeek | 90.9 | Level I accuracy % | zero-shot | Reported by the authors | Paper Table 5 | https://arxiv.org/html/2512.08270v1 | 9 Dec 2025 (checked 1 Nov 2026) |

## TaxEval v2: accuracy % (v2) (higher is better, 5 models)

| Rank | Model | Provider | Score | Metric | Setting | Reported by | Source | URL | Date |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| 1 | Muse Spark 1.2 | Meta | 80.4 | accuracy % (v2) | test set (1,223 questions), archived board | Official board | Vals.ai TaxEval v2 leaderboard | https://www.vals.ai/benchmarks/tax_eval_v2 | 1 Sept 2026 (checked 2 Nov 2026) |
| 2 | Muse Spark 1.1 | Meta | 79.7 | accuracy % (v2) | test set (1,223 questions), archived board | Official board | Vals.ai TaxEval v2 leaderboard | https://www.vals.ai/benchmarks/tax_eval_v2 | 1 Sept 2026 (checked 2 Nov 2026) |
| 3 | Muse Spark | Meta | 77.7 | accuracy % (v2) | test set (1,223 questions), archived board | Official board | Vals.ai TaxEval v2 leaderboard | https://www.vals.ai/benchmarks/tax_eval_v2 | 1 Sept 2026 (checked 2 Nov 2026) |
| 4 | Claude Sonnet 4.6 | Anthropic | 77.1 | accuracy % (v2) | test set (1,223 questions), archived board | Official board | Vals.ai TaxEval v2 leaderboard | https://www.vals.ai/benchmarks/tax_eval_v2 | 1 Sept 2026 (checked 2 Nov 2026) |
| 5 | Claude Fable 5 | Anthropic | 76.9 | accuracy % (v2) | test set (1,223 questions), archived board | Official board | Vals.ai TaxEval v2 leaderboard | https://www.vals.ai/benchmarks/tax_eval_v2 | 1 Sept 2026 (checked 2 Nov 2026) |

## τ³-Banking: pass^1 % (higher is better, 5 models)

| Rank | Model | Provider | Score | Metric | Setting | Reported by | Source | URL | Date |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| 1 | Qwen 3.8 Max | Qwen | 55.2 | pass^1 % | all tools retrieval, gpt-5.2 user simulator | Official board | τ-bench leaderboard (Sierra) | https://taubench.com/leaderboard/?benchmark=knowledge | 4 Aug 2026 (checked 2 Nov 2026) |
| 2 | Claude Opus 5 | Anthropic | 48.7 | pass^1 % | all tools retrieval, gpt-5.2 user simulator | Official board | τ-bench leaderboard (Sierra) | https://taubench.com/leaderboard/?benchmark=knowledge | 4 Aug 2026 (checked 2 Nov 2026) |
| 3 | Grok 4.5 | xAI | 47.9 | pass^1 % | all tools retrieval, gpt-5.2 user simulator | Official board | τ-bench leaderboard (Sierra) | https://taubench.com/leaderboard/?benchmark=knowledge | 4 Aug 2026 (checked 2 Nov 2026) |
| 4 | GPT-5.6 Sol | OpenAI | 46.9 | pass^1 % | all tools retrieval, gpt-5.2 user simulator | Official board | τ-bench leaderboard (Sierra) | https://taubench.com/leaderboard/?benchmark=knowledge | 4 Aug 2026 (checked 2 Nov 2026) |
| 5 | GPT-5.5 | OpenAI | 44.6 | pass^1 % | all tools retrieval, gpt-5.2 user simulator | Official board | τ-bench leaderboard (Sierra) | https://taubench.com/leaderboard/?benchmark=knowledge | 5 May 2026 (checked 2 Nov 2026) |

## Also in this use case (no dated scores yet)

- FinBen (The Fin AI (Qianqian Xie, Jimin Huang et al., 34 authors)): https://benchmarks.finance/benchmarks/finben/
- Moneda KnowledgeBench (Moneda): https://benchmarks.finance/benchmarks/moneda-knowledgebench/

Source: https://benchmarks.finance/best/finance-knowledge/

Created and maintained by Moneda.