Vals AI Finance Agent v2.1: leaderboard
Finance Agent v2.1: 449 private test questions across nine financial-analysis categories, dealbreaker-gated partial credit. Revised judge jury, hedged-answer prompt and expert-reviewed questions/rubrics; not comparable with v2. 27 public examples and 450 licensed validation questions are separate.
Metric: Dealbreaker-gated Partial Credit (%). Source: www.vals.ai. Saturation forecast: Around 2030. 28 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 5.5 (Max) | 67.85 | #4 (Claude Opus 5.5) |
| 2 | Gemini 4 Argon (High) | 67.66 | #7 (Gemini 4 Argon) |
| 3 | Muse Wonder Full (Max) | 67.58 | #10 (Muse Wonder Full) |
| 4 | Claude Sonnet 5.5 (Max) | 67.28 | #11 (Claude Sonnet 5.5) |
| 5 | Gemini 3.8 Flash (High) | 65.76 | #23 (Gemini 3.8 Flash) |
| 6 | Muse Wonder (Max) | 65.55 | #14 (Muse Wonder) |
| 7 | DeepSeek V4.1 Flash (Max) | 65.43 | #36 (DeepSeek V4.1 Flash) |
| 8 | Claude Haiku 5.5 (Max) | 65.37 | #40 (Claude Haiku 5.5) |
| 9 | GLM-5.3 (Max) | 65 | #39 (GLM-5.3) |
| 10 | MiMo-V2.6-Pro | 64.26 | #42 |
| 11 | MiMo-V2.6-Flash | 63.32 | #71 |
| 12 | DeepSeek V4.1 Flash (1017) (Max) | 62.93 | #20 (DeepSeek V4.1 Flash (1017)) |
| 13 | GPT-6.1 Sol (Max) | 61.82 | #8 (GPT-6.1 Sol) |
| 14 | Hy4 preview | 61.62 | #38 |
| 15 | Mistral Large 4 (High) | 61.38 | #95 (Mistral Large 4) |
No result here: #9 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=vals-ai-finance-agent-v2-1 · How It Works · Data refreshed daily, snapshot 2026-10-21.