Unverified50% confidenceFactExact time
GAIA由Meta提出，专门评测Agent的真实世界问答和工具使用能力，要求通过多步骤搜索、计算和推理得出答案
1
Sources
50%
Confidence
Long-term
Relevance
8/5/2026
First Seen
Sources
Related Claims
Unverified现有基准GAIA、BrowseComp只提供明确查询，仅考验多步推理能力，未考虑歧义问题63% similarUnverifiedAgentic benchmark evaluations typically involve a weighted aggregation of multiple sub-benchmarks such as SWE-bench, WebArena, and GAIA.63% similarUnverifiedRAG (Retrieval-Augmented Generation) knowledge bases are one of the most frequently asked-about technical topics in AI-empowered testing interviews.61% similarUnverifiedEvaluation of deep research agents can reference the RAGAS framework, assessing across multiple dimensions such as faithfulness and answer relevance.61% similarUnverified语义搜索或推荐系统背后运用的是嵌入技术61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/688907API
curl https://kongchang.com/api/v1/knowledge/claims/688907MCP
get_claim(id=688907)