Super-adapter for Qwen3.8-27B: concatenated LoRAs beat the thinking model
yb2698 · x · 2026-11-20
yb2698 announces an experimental "super-adapter" for Qwen3.8-27B: a bigger LoRA that merges all harness adapters from the week via a simple trick — concatenating all LoRA adapters into one. It beats the thinking model while making 19–48% fewer tool calls in thinking mode.
Results: SWE-Bench Pro (hard 51) — Hermes Agent 45%→61%, OpenCode 47%→67%, pi 74%→74%; Terminal-Bench 2.1 — +4 to +6 points per harness. The full recipe, experimental results, and open questions are in the accompanying blog post.
More from coding & agent
- Grok Bot Gets Its Own Email Addresses to Sign Up and Handle Promos for You — elonmusk · 2026-11-21
- CapCut × ChatGPT plugin turns chat into a video generator and editor — SimplyAnnisa · 2026-11-21
- Dev builds daily X-digest Grok bot that auto-blocked 19 reply bots, heading to marketplace — petergyang · 2026-11-21
- DeepMind's Antigravity Robotics Plugin: 2 Prompts, ~20 Minutes to Command a Real FR3 Duo Arm — jackyliang42 · 2026-11-21
- Asking Astra 6.2 to Review Code Written by Opus 5.5 Yields Almost Nothing — rickasaurus · 2026-11-21
- Security & Web3 Firewall MCP Server Hits Official Registry: Detects Tool-Poisoning and Rug-Pull Drift — ezelech · 2026-11-21