Hands-on with Qwen 3.8: Extremely capable but not GPT 5.6

gnukeith · x · 2026-08-19

The author shared their experience locally hosting the Qwen 3.8 model. While not as hyped as Fable or GPT 5.6, it remains an extremely capable and smart model. Its conversational style is reminiscent of Kimi K2.5—optimistic and grounded—unlike other models that feel mechanical and dismissive.

Original post →

More from Models

Models channel →