Unverified50% confidenceFactExact time
FrontierSWE 由 METR（Model Evaluation and Threat Research）组织推出，衡量智能体完成数小时到数十小时规模的开放式技术项目
1
Sources
50%
Confidence
Long-term
Relevance
8/27/2026
First Seen
Sources
Related Entities
Related Claims
VerifiedFrontierCode is an evaluation benchmark for AI models' programming capabilities focusing on real engineering tasks rather than simple algorithm problems63% similarUnverifiedOrganizations such as METR (Model Evaluation and Threat Research) are developing standardized dangerous capability evaluations, including whether a model can provide biosynthesis guidance beyond publicly available literature and whether it can autonomously discover zero-day vulnerabilities.62% similarUnverifiedAnthropic uses the RSP framework to evaluate dangerous capability thresholds in models, OpenAI uses the Preparedness Framework, and Google DeepMind uses the Frontier Safety Framework62% similarUnverifiedThe core philosophy of Frontier Code is to evaluate whether code can be merged by project maintainers (PR mergeability), rather than merely passing tests.60% similarUnverifiedMETR (formerly ARC Evals) focuses on evaluating the dangerous capabilities of frontier models58% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/807814API
curl https://kongchang.com/api/v1/knowledge/claims/807814MCP
get_claim(id=807814)