How one dev routes coding tasks across 5 model tiers from local Qwen to Opus
Short_Regular_7191 · reddit · 2026-10-12
A solo maintainer of a large production web app (Node/Express, Svelte, PostgreSQL+Oracle, hundreds of Cypress tests) shares his multi-tier coding stack: local Qwen3.8 Coder (free, no-decision work) → Haiku 5.5 (parallel mechanical lanes) → Sonnet 5.5 (orchestration, diff review) → Opus 5.5 (parent session, debugging, DB ops) → Fable 5.1 (strict audits and architecture). The biggest token savings come from keeping the orchestrator's context small: sub-agents read big files and reply with 5-line summaries, repeated work becomes scripts the model only reads output from, context compacts at 80%, and models are never switched mid-session to preserve caching. Tasks are graded on an R0–R9 matrix, with DB/irreversible operations restricted to Opus in-session plus human approval.
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
- 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
- Let agents design a single-line glance component so you can track them from the sidebar — blixt · 2026-10-23