Tsinghua open-sources PilotDeck 2.0: agent OS cuts LLM costs 83% with smart routing
量子位 · wechat · 2026-10-17
Tsinghua THUNLP, ModelBest and OpenBMB released PilotDeck 2.0, an open-source (AGPL-3.0) agent OS tackling memory management, task execution, and cost.
Key features: WorkSpace-based project isolation, fully white-box memory (inspectable, editable, with Dream-mode background consolidation and rollback), Always-on background task execution, native MCP support across Web/CLI/IM, and Office file processing with checkpoint resume.
Its "strong-main, weak-sub" smart routing scores 70.6 at $3.15 versus single-model Sonnet 4.6's 69.1 at $18.36 — an 83% cost cut. On Workbuddy Bench it scores 79.1 with Qwen3.6-35B-A3B, beating Claude Code's 77.3.
More from coding & agent
- OpenJet claims Qwen3.8-27B coding runs 1.9x faster than llama.cpp on one RTX 3090 — JeremyCMorgan · 2026-10-23
- Jensen Huang details Vera Rubin agentic stack: NVL72, MGX, CMX and RTX — rohanpaul_ai · 2026-10-23
- Building an autonomous SEO agent in Hermes with three skills: programmatic pages, Keywords Everywhere, DataForSEO — EXM7777 · 2026-10-23
- Building a sports betting AI agent with onchain execution and human approval by default: architecture and lessons — Dry-Comfortable-2678 · 2026-10-23
- Capy 0.5.0 lets Claude Pro/Max subscribers reuse their plan quota, adds OIDC cloud access — garrytan · 2026-10-23
- You.com CEO Richard Socher: legacy code is human slop, let AI clean it up — RichardSocher · 2026-10-23