Dev pushes Qwen 27B to 70-90 tok/s on a 4090 in a local AI coding workflow
julianharris · x · 2026-09-20
Developer julianharris shares his fast local AI coding setup:
- Tuning win: a setting bumped 4090 + Qwen 27B from 50-70 to sustained 70-90 tok/s — roughly 2x Claude Opus throughput
- Stack: the minimal Pi agent harness plus an MCP extension wired into his spec management/governance system (Ceetrix)
- Case study: Qwen found a bug that looked like it required a rewrite, but Ceetrix task context revealed the feature existed — just behind a menu instead of a button
- Pi: supports skills, AGENTS.md, four modes (interactive/print/RPC/SDK), and self-customization via extensions
More from coding & agent
- minpts-check: one-run predictor tells if your HDBSCAN clusters survive min_samples changes — leland_mcinnes · 2026-11-12
- Ten wild Gemini 4 Carbon demos in 24 hours: CUDA meadows, V8 engines, mech stomps — socialwithaayan · 2026-11-12
- How to cut Fable 5.2 token costs by 90%: Arc routing plus a lean CLAUDE.md — EXM7777 · 2026-11-12
- Hamel Husain's complete AI evals framework for PMs and engineers — aakashgupta · 2026-11-12
- AI coding in one year: from unusable to full automation with no code review — angelaqdai · 2026-11-12
- Interactions Tool Searches Web Animations by Motion for Claude Code to Rebuild — yuwen_lu_ · 2026-11-12