Qwen3.8-27B: 27B params rival frontier, beats Claude Opus 4.6 Max on agent benchmarks?

kimmonismus · x · 2026-08-14

X user kimmonismus comments on Qwen3.8-27B release, claiming it surpasses Claude Opus 4.6 Max on agentic coding and computer use benchmarks despite 27B params, with 262K context and Apache 2.0. Notes benchmarks are self-reported.

Original post →

More from Models

Models channel →