FinRiskAtlas - Applied Review - Disputed-Issue Generation: leaderboard
Open-ended FinRiskAtlas review generation: writing a structured list of contested issues in a case with overlapping legal and contractual relationships, scored by a fixed semantic rubric for correctness, coverage and evidence use with DeepSeek-V4-Flash as evaluator (FinRiskAtlas, arXiv 2608.25325v1, Appendix Table 10). Percent; higher is better. The evaluator's own self-scored row is excluded, as in the paper's comparisons.
Metric: Rubric Score (%; DeepSeek-V4-Flash judge). Source: arxiv.org. Saturation forecast: Around December 2027. 32 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | DeepSeek V4 Pro | 77.12 | #112 |
| 2 | Nemotron 3 Super | 76.67 | #344 |
| 3 | Kimi K3 | 74.84 | #35 |
| 4 | Gemini 3 Flash (Preview) | 74.69 | #90 |
| 5 | Claude Opus 4.7 | 74.06 | #51 |
| 6 | Gemini 2.5 Flash | 74.05 | #255 |
| 7 | Seed 2.0 Lite | 73.35 | #161 |
| 8 | Seed 2.0 Mini | 72.73 | #279 |
| 9 | GLM-5 | 72.49 | #149 |
| 10 | Nemotron 3 Ultra | 71.58 | #245 |
| 11 | Qwen 3.7 Max | 70.58 | #85 |
| 12 | Qwen 3 4B | 68.83 | #834 |
| 13 | GLM-5.2 | 68.37 | #80 |
| 14 | Qwen3-Next-80B | 67.51 | #321 |
| 15 | Ling-2.6-1T | 63.3 | #307 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=finriskatlas-applied-review-disputed-issue-generation · How It Works · Data refreshed daily, snapshot 2026-11-05.