AI BriefingAnthropic Engineering news
The latest AI and developer news from Anthropic Engineering, with short summaries.
Visit website →How We Built a Multi-Agent Research System
Parallel subagents and prompt/eval design pushed Research's performance higher.
Claude Code: Best Practices for Agentic Coding
Claude Code is an agentic coding tool where verification, planning, and context management are key.
The "think" tool that makes Claude pause and reason in complex tool-use situations
With the think tool, Claude improves policy compliance and consistency in complex tool use.
Raising the Bar on SWE-bench Verified with Claude 3.5 Sonnet
The upgraded Claude 3.5 Sonnet scored 49% on SWE-bench Verified.
Building Effective Agents
The best LLM systems come not from complex frameworks but from simple patterns.
Introducing Contextual Retrieval
Attaching document context before and after chunks to build embeddings and BM25 can significantly reduce RAG retrieval failures.
Update on Recent Claude Code Quality Reports
Anthropic said it identified and fixed the cause of Claude Code quality degradation as 3 separate changes.
Quantifying Infrastructure Noise in Agentic Coding Evaluations
How infrastructure resource configuration is set during agentic coding evaluations has a significant impact on the measured results of model performance.