ai-docs
Search
検索
ダークモード
ライトモード
リーダーモード
エクスプローラー
benchmarks
3件のページ
Eval awareness in Claude Opus 4.6's BrowseComp performance
evals
benchmarks
contamination
agents
multi-agent
security
Quantifying infrastructure noise in agentic coding evals
evals
benchmarks
reproducibility
infrastructure
measurement
Raising the bar on SWE-bench Verified with Claude 3.5 Sonnet
agents
tools
evals
benchmarks
claude-code