Unverified50% confidenceFactExact time
Agent's Last Exam 是专为评估大模型自主代理能力设计的高难度基准，要求多步骤任务中持续调用外部工具、维持长上下文记忆并自我修正
1
Sources
50%
Confidence
Long-term
Relevance
7/13/2026
First Seen
Sources
GPT-5.6三档模型解析：Sol/Terra/Luna重写Agent成本结构
bilibiliTerminator-AI7/10/2026
Related Claims
UnverifiedAgent's Last Exam 是由 Scale AI 与学术界联合设计的 Agent 能力基准测试83% similarUnverifiedAgents' Last Exam is an evaluation system covering 55 professional domains that simulates long-running professional workflows.79% similarUnverifiedKey capabilities of Autonomous Agents include Tool Use, Memory Management (short-term and long-term), and Self-Reflection for evaluating and correcting execution strategies.75% similarUnverifiedAn Agent consists of a Planning module, a toolset, and short-term and long-term memory, ultimately outputting Action instructions to complete invocations.75% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/501596API
curl https://kongchang.com/api/v1/knowledge/claims/501596MCP
get_claim(id=501596)