AI EVALUATION, IN CONTEXT.BY EQUALITY MACHINE

Model reports

GDPval

Graded by expert human comparison against real deliverables, and the "-AA" variants are run by Artificial Analysis rather than the vendor, so an Elo here is not comparable to a vendor-run pass rate.

评测信息与证据

来源
Model reports
发布者
未收录
评测发布日期
2025-09-25
采集日期
未收录
模态
未收录

本站保留来源报告的结果,不代表独立复测。请先核对任务版本、评测协议、分数单位和报告日期;不同运行条件不能直接合并为统一排名。

原始资料

报告成绩 6

按来源中的观察顺序展示,顺序不代表排名。日期保留原始精度。

模型数值单位协议报告日期与口径来源
Claude Opus 51861.0eloELO rating, max effort, agentic loop with shell access and web browsing2026-07-24 · document_publicationClaude Opus 5
DeepSeek-V4-Pro1554.0elothink max, Elo2026-04-22 · document_publicationDeepSeek-V4-Pro
DeepSeek-V4-Flash1395.0elothink max, Elo2026-06-25 · document_publicationDeepSeek-V4
Kimi K31686.0eloElo, reasoning=max2026-06-13 · document_publicationKimi K3
Hy4 preview1678.0eloGDPval-AA V2, Elo, official2026-08-28 · document_publicationHy4 preview
GLM-5.3-Flash1773.0eloGDPval-AA v2, evaluated by Artificial Analysis2026-08-27 · document_publicationGLM-5.3-Flash