SOURCE-REVIEWED MODEL PROFILE
Qwen3.8-27B super-adapter
Merged adapter for three coding-agent harnesses; not a standalone foundation model.

Metadata reviewed 2026-11-20 · Original creator: harness-hub
What it does
- Three rank32 adapters concatenated at half weight
- Publisher recommends thinking mode for merged use
- Base checkpoint required
Local fit
Requires base-model memory plus adapter/context overhead;not tested.. Reported benchmark benefits depend on harness and thinking mode. No clear general business-automation transfer gain.
Video streaming uses a rolling cache;agent adapter context depends on the base and runtime. No independent long-context test.
Evidence and limits
Source reviewed;no local inference,hardware fit or benchmark test.
Publisher reference workflow;not independently tested
License and source
Model card declares Apache-2.0; no standalone LICENSE file verified in this artifact. Base and runtimes separately licensed.
Read the original model card ↗
Reviewed source revision: 3358a71529ce56aba0095addd6a60a66d7c4228b