Stackness9 min read
Does a harness make a cheap model as reliable as a frontier model? Harnessed GPT-6 Luna against bare Sol
In a test published 13 October 2026, GPT-6 Luna in a harness passed 10 of 10 like bare Sol, at $0.273 a run against $0.406, but took 5x as long.
By Sergei GordeichukStackness Blog

Read the full story
Originally published by Stackness Blog on Oct 23, 2026.
Comments
Sign in to join the discussion.