Best AI for finance (Nov 2026)
Analyst research on SEC filings and long credit agreements. Pick one of 2 leaderboards; the leaders on each, with sources, are summarised below the chart.
Finance Agent
Entry-level financial analyst work: answering difficult questions on public company filings.
Loading chart…
Each square is one model, marked with its provider’s initial. Higher is better; further left costs less per test.
Highest reported score: Claude Fable 5.2 (Anthropic) at 69.47%.
Accuracy, highest first. Position is the source’s published rank.
- 1Claude Fable 5.2 Anthropic69.47%
- 2Claude Opus 5.5 Anthropic67.85%
- 3Gemini 4 Argon Google67.66%
- 4Wonder Full Meta67.58%
- 5Inkling 2 Small Other67.54%
- 6Claude Sonnet 5.5 Anthropic67.28%
- 7Gemini 4 Carbon Google67.16%
- 8Qwen 4 Max Alibaba66.68%
- 9Gemini 3.8 Flash Google65.76%
- 10Wonder Meta65.55%
- 11DeepSeek V4.1 Flash DeepSeek65.43%
- 12Claude Haiku 5.5 Anthropic65.37%
- 13Gemini 4 Hydrogen Google65.12%
- 14GLM 5.3 Z.ai65.00%
- 15Hy4 Other64.70%
- 16MiMo V2.6 Pro Xiaomi64.26%
- 17Muse Ember Meta63.83%
- 18MiMo V2.6 Flash Xiaomi63.32%
- 19DeepSeek V4.1 Flash 1017 DeepSeek62.93%
- 20GPT-6.1 Sol OpenAI61.82%
- 21Mistral Large 4 Mistral61.72%
- 22Hy4 Preview Other61.62%
- 23Kimi K3.1 Pro Moonshot61.49%
- 24Mistral Large 4 Preview Mistral61.38%
- 25Muse Spark Open Meta60.66%
| 1 | Claude Fable 5.2 | Anthropic | 69.47% | $4.83 |
|---|---|---|---|---|
| 2 | Claude Opus 5.5 | Anthropic | 67.85% | $8.91 |
| 3 | Gemini 4 Argon | 67.66% | $4.32 | |
| 4 | Wonder Full | Meta | 67.58% | $2.68 |
| 5 | Inkling 2 Small | Other | 67.54% | $0.18 |
| 6 | Claude Sonnet 5.5 | Anthropic | 67.28% | $6.95 |
| 7 | Gemini 4 Carbon | 67.16% | $7.29 | |
| 8 | Qwen 4 Max | Alibaba | 66.68% | $1.74 |
| 9 | Gemini 3.8 Flash | 65.76% | $1.99 | |
| 10 | Wonder | Meta | 65.55% | $2.62 |
| 11 | DeepSeek V4.1 Flash | DeepSeek | 65.43% | $0.37 |
| 12 | Claude Haiku 5.5 | Anthropic | 65.37% | $1.64 |
| 13 | Gemini 4 Hydrogen | 65.12% | $4.36 | |
| 14 | GLM 5.3 | Z.ai | 65.00% | $1.03 |
| 15 | Hy4 | Other | 64.70% | $1.28 |
| 16 | MiMo V2.6 Pro | Xiaomi | 64.26% | $0.18 |
| 17 | Muse Ember | Meta | 63.83% | $0.060 |
| 18 | MiMo V2.6 Flash | Xiaomi | 63.32% | $0.070 |
| 19 | DeepSeek V4.1 Flash 1017 | DeepSeek | 62.93% | $0.47 |
| 20 | GPT-6.1 Sol | OpenAI | 61.82% | $1.63 |
| 21 | Mistral Large 4 | Mistral | 61.72% | $0.35 |
| 22 | Hy4 Preview | Other | 61.62% | $0.67 |
| 23 | Kimi K3.1 Pro | Moonshot | 61.49% | $2.29 |
| 24 | Mistral Large 4 Preview | Mistral | 61.38% | $0.52 |
| 25 | Muse Spark Open | Meta | 60.66% | $2.76 |
Source: Vals AI · leaderboard dated 11 Nov 2026 · checked 17 Nov 2026
Open the full Finance Agent leaderboard
Prompts for the work Finance Agent tests
All prompts →Results from published AI leaderboards. Compare models within each benchmark; scores from different tests aren’t directly comparable. How we collect these results
Highest reported scores
As of 11 Nov 2026, the highest reported score on Finance Agent is Claude Fable 5.2 (Anthropic) at 69.47%, per Vals AI. As of 12 Aug 2026, the highest reported score on CorpFin v2 is Claude Opus 5 (Anthropic) at 73.19%, per Vals AI. The source no longer updates this leaderboard. No single model has the highest reported score on every finance leaderboard here, so compare within each benchmark.
All 2 finance benchmarks
Finance Agent
Entry-level financial analyst work: answering difficult questions on public company filings.
As of 11 Nov 2026, the highest reported score on Finance Agent is Claude Fable 5.2 (Anthropic) at 69.47%, per Vals AI.
CorpFin v2
Understanding long-context credit agreements (private benchmark).
As of 12 Aug 2026, the highest reported score on CorpFin v2 is Claude Opus 5 (Anthropic) at 73.19%, per Vals AI. The source no longer updates this leaderboard.
Questions people ask
Which AI has the highest reported scores for finance?
As of 11 Nov 2026, the highest reported score on Finance Agent is Claude Fable 5.2 (Anthropic) at 69.47%, per Vals AI. As of 12 Aug 2026, the highest reported score on CorpFin v2 is Claude Opus 5 (Anthropic) at 73.19%, per Vals AI. The source no longer updates this leaderboard. No single model has the highest reported score on every finance leaderboard here, so compare within each benchmark.
Which benchmarks are on this page?
Finance Agent (run by Vals AI): Entry-level financial analyst work: answering difficult questions on public company filings. CorpFin v2 (run by Vals AI): Understanding long-context credit agreements (private benchmark).
How recent are these results?
Finance Agent: leaderboard dated 11 Nov 2026; we last checked it on 17 Nov 2026. CorpFin v2: leaderboard dated 12 Aug 2026, no longer updated by the source; we last checked it on 17 Nov 2026.
Does ChatJudge run these tests?
No. Every score is copied as published by the leaderboard that ran it, with a link to the source. Scores from different benchmarks are not directly comparable.
