Touchstone
← Harnesses

v1.7.1

localharnesscurrent

Per-domain thinking control: reasoning models run NO-THINK (prefilled empty think block) on gate + coding/C#/PHP ladders, and keep FULL thinking on SupportFlow.

why it exists: Reasoning models over-think toy benches and time out; thinking earns its keep only on the hard agentic domain. Split measures real capability.

Models
1
have run it
Solved
35
gauntlet benches

Models on this harness (gauntlet solved / attempted)

ModelSolvedBy domain
qwen3.6-27b35/37
Coding 6/6C# 13/15Gate 11/11PHP 5/5