Claude Opus · Finance Agent · Vals AI SEC-filing research benchmark · 历代记录

按发布时间查看 Claude Opus 的 2 次发布在 Finance Agent · Vals AI SEC-filing research benchmark 的成绩、配置与披露。

仅为直接前代间的唯一、有效精确值计算差值,要求单位、指标与明确记录的配置和条件一致。

Claude Opus 4.5

· 本次发布有 1 条成绩记录,含 1 条精确值(可含历史记录)

Claude Opus 4.5 · Finance Agent 成绩与配置
指标与成绩配置与评测条件出处
Accuracy55.2%原文 55.2% · 单位 %
公布 未确认
FinanceAgent external Vals AI evaluation
设置
64K thinking budget; 200K context; external Vals AI analysis; eight-trial average.
未知项
The system card does not state a named FinanceAgent version or effort label.
工具
工具条件未记录
推理预算
思考预算 64K
框架
未记录
协议
FinanceAgent external Vals AI analysis; version not stated; eight-trial average.
条件
64K thinking budget, 200K context. The two higher Anthropic self-evaluation results are internal tests and are excluded.
在排名中查看此成绩

Claude Opus 4.7

· 本次发布有 1 条成绩记录,含 1 条精确值(可含历史记录)

Claude Opus 4.7 · Finance Agent 成绩与配置
指标与成绩配置与评测条件出处
Accuracy64.4%原文 64.4% · 单位 %
公布 未确认
发布表报告值
设置
System Card §8.1 standard configuration: adaptive thinking at max effort; default temperature/top_p; averaged over 5 trials. Context sizes are evaluation-dependent and do not exceed 1M tokens; retain all section-specific exceptions on their results.
未知项
Per-benchmark tools, context/token cap, resolution, trial count, harness and other exceptions are recorded on each result when stated.
工具
工具条件未记录
推理预算
high
框架
未记录
协议
Finance Agent v1.1; Vals AI; adaptive thinking, high effort。
条件
Finance Agent v1.1; Vals AI evaluation using adaptive thinking and high effort. The card body prints a duplicated percent sign (64.4%%); the table and announcement report 64.4%.
在排名中查看此成绩