FinReasoning (Data Alignment) - Verification: leaderboard
Metric: Success-weighted score (0-100): the success rate (items whose query ran and whose answer came back in the required format) times the mean of answer accuracy, F1 over the cited data IDs and F1 over the cited fields, on the 600 verification items (whether a perturbed statement matches the database) of FinReasoning's Data Alignment track (answers checked against a structured A-share database by rule); Chinese-language tasks built from A-share research reports, news and market data (2023 to 2025); zero-shot at temperature 0.1; higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 19 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Pro | 94.3 | #77 |
| 2 | Seed 1.8 | 93.5 | #136 |
| 3 | Qwen 3 Max | 92.3 | #201 |
| 4 | DeepSeek V3 | 91.8 | #312 |
| 5 | Qwen 3 32B | 91.6 | #424 |
| 6 | GPT-4o | 90.4 | #333 |
| 7 | Kimi K2 | 90.2 | #236 |
| 8 | Qwen 3 235B A22B | 90.1 | #304 |
| 9 | GLM-4.6 | 89.4 | #246 |
| 10 | Intern-S1 | 88.8 | #278 |
| 11 | GPT-5 | 88.3 | #91 |
| 12 | Llama 3.1 70B | 88.2 | #578 |
| 13 | Claude Sonnet 4.5 | 85.3 | #138 |
| 14 | Qwen 3 8B | 83.5 | #667 |
| 15 | Llama 3.1 8B | 78.4 | #1139 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=finreasoning-data-alignment-verification · How It Works · Data refreshed daily, snapshot 2026-10-11.