Finance Agent v1.1: leaderboard
537 entry-level financial-analyst questions (data retrieval, market research, projections) answered with tool calls; Vals AI with Stanford researchers and a global systemically important bank.
Metric: Accuracy (%). Source: www.vals.ai. Status: saturation imminent. 51 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 4.7 (High) | 64.37 |
| 2 | Claude Sonnet 4.6 (Max) | 63.33 |
| 3 | GPT-5.4 Pro (xHigh) | 61.5 |
| 4 | Muse Spark | 60.59 |
| 5 | DeepSeek V4 Pro (Max) | 60.39 |
| 6 | Claude Opus 4.6 | 60.1 |
| 7 | Claude Opus 4.6 (Adaptive Reasoning, Max Effort) | 60.05 |
| 8 | GPT-5.5 (xHigh) | 59.96 |
| 9 | Gemini 3.1 Pro (Preview) (High) | 59.72 |
| 10 | GPT-5.2 (xHigh) | 58.53 |
| 11 | GLM-5.1 | 57.66 |
| 12 | GPT-5.4 (xHigh) | 57.15 |
| 13 | Kimi K2.6 | 57.06 |
| 14 | GPT-5.1 (High) | 55.31 |
| 15 | Gemini 3 Pro (Preview) (High) | 55.15 |
Interactive version: theaggregate.ai/benchmark?slug=finance-agent-v1-1 · How It Works · Data refreshed daily, snapshot 2026-09-05.