FinRiskAtlas - FinRisk-Ask - Balanced Action Agreement: leaderboard

Balanced agreement (BAcc) with recorded reviewer Ask-or-Proceed decisions on 680 replayed FinRisk-Ask review states, the mean of Ask recall and Proceed recall (FinRiskAtlas, arXiv 2608.25325v1, Appendix Table 11). Percent; higher is better; constant always-ask or always-proceed policies score 50. It measures agreement with one recorded workflow, not a unique optimal policy.

Metric: BAcc (%). Source: arxiv.org. 33 models tracked.

Top models

#ModelScoreOverall rank
1Kimi K359.46#35
2Qwen 3 8B59.39#678
3Qwen 3 32B58.19#449
4Qwen3-Next-80B56.21#321
5Claude Opus 4.756.11#51
6GPT-5.6 Terra56.11#40
7Ling-2.6-flash56.03#533
8Nemotron 3 Ultra55.69#245
9Kimi K2.555.68#152
10GLM-555.26#149
11Qwen 3 14B54.83#542
12Qwen 3 4B54.68#834
13GPT-5.6 Luna54.21#79
14GPT-5.6 Sol54.05#22
15Qwen 3 235B A22B53.63#323

Interactive version: theaggregate.ai/benchmark?slug=finriskatlas-finrisk-ask-balanced-action-agreement · How It Works · Data refreshed daily, snapshot 2026-11-05.