FLock ships 1.88B THIS/THAT model 1.2, claims 87.8% on its own decision test

matlabulous · x · 2026-09-24

FLock released THIS / THAT model 1.2, a tiny 1.88B-parameter model that generates zero tokens and scored 87.8% on the company's own "super complex decision test" — up from 40.6% (v1.0) and 77.5% (v1.1) — claiming it now beats Claude Opus 5 and GPT-5.6 on that benchmark. The model is free via the FLock API. Note the comparison is on a proprietary in-house test, not public benchmarks.

Original post →

More from Models

Models channel →