Anthropic's 'Graph Engineering' Leaks: Multi-Agent Crews Cut Costs 46%
anirbanbandyo · x · 2026-08-15
A leak claims Anthropic uses 'Graph Engineering' to run AI crews instead of single models. The architecture assigns roles: one planner, several executors, and one fact-checker. Their internal metrics show this setup maintains 96% of the output quality while reducing costs by 46%. This suggests a shift towards multi-agent orchestration as the next revolution in AI workflow.
More from coding & agent
- Steve Yegge's Beads agentic memory layer hits 1.5M downloads one year after launch — doodlestein · 2026-10-16
- Sakana AI open-sources AI-Scientist-v3, orchestrating coding agents to automate scientific discovery — cong_ml · 2026-10-16
- Coding agents keep opening 3 stacked PRs per chat, and merging them is a nightmare — holdenmatt · 2026-10-16
- Google-led paper uses sequential auctions to organize decentralized LLM agent workflows — _onionesque · 2026-10-16
- Since OpenAI's math release, GitHub contributors keep driving down the integer-multiplication kappa — xennygrimmato_ · 2026-10-16
- Pattern MCP makes coding agents check existing components before writing new code — DonR954 · 2026-10-16