Qwen 3.8 27b on the agent mode harness.
Built itself a python app to generate the single html. Prompt in the first comment proves the harness value 🏆👇⬇️ ..also impressive workspace structure in agent mode 👏
ABSURD. Qwen 3.8 27B is cruising locally on an RTX 4060 with just 8GB of VRAM.
That’s a 64k context window courtesy of Unsloth’s brand-new IQ4_XS quant, and the whole thing weighs in at 14.6GB on disk. Prefill lands around 150 tok/s; decode hovers at ~5 tok/s with native MTP.
Japanese Tea Garden Bench - Qwen 3.8 Flash Next
Just lucky, I guess. This one took 6 hours and several million tokens, at medium effort. I would say good job, but the z-fighting is prominent, and it might never have finished had I not stopped it.
This is my personal experience with Qwen 3.8 27B in Hermes versus Grok 4.6 in Grok Build.
You have reached the end of the archive
All of qwen38