Some quick Qwen 3.8 demos since our time with 3.8 is almost over.
Strata IQ3_S - Qwen 3.8 Flash Next and Qwen 3.8 Flash Next FP8 official Ran both locally, ran my little colosseum prompt through both using Deepseek harness, pretty good results on both.
What if AI agents stopped doing all the work and started managing it instead?
👀 Fable 5.2 was tested with 3 local Qwen 3.8 27B agents handling the actual building. The results: ⚡ 21 min → 11 min 💰 $15.77 → $2.98 🤖 2,860 lines of code written by Qwen 🧠 Fable focused on
Welcome to my small pond of pokemons commonstack_ai
Qwen 3.8 flash is quite elite
Qwen 3.8 27B on my Mac Studio was the only model that actually built the glacier, the aurora and the title from my prompt, at 61 tok/s.
Same prompt on three stacks. GLM 5.3 Flash on the 4 box Spark ring hit 180 tok/s and DeepSeek V4.1 Flash hit 150. Speed isn't everything.
You have reached the end of the archive
All of qwen38