Cheers for the new local GOAT!

Also on ModelScope in case you-know-who bribes huggingface takedown because model is too dangerous.

Qwen3.6 with 4-bit quant was hallucinating HARD. When asked to summarize the plot of obscure, cheesy 80s books, it confidently made up plot and characters. Not a single statement was correct. It was glorious.

Any higher quant and it would take 30-90 seconds before first token. Both MLX and GGUF on an M3Max with 128G combined RAM. Will have to try it with this one.

Has anyone here really pushed the multimodal stuff with this yet? Been running this on a 4090 for over 24 hours and have been meaning to try out if its able to summarize images well, especially compared to something like qwen3vl on the same hardware

Having issues with it (Unsloth's gguf).

0.27.136.404 I srv  proxy_reques: proxying request to model unsloth/Qwen3.8-27B-GGUF:Q4_K_M on port 55741
[55741] 0.14.424.772 W srv    operator(): got exception: {"error":{"code":500,"message":"\n------------\nWhile executing CallExpression at line 106, column 32 in source:\n...first %}↵            {{- raise_exception('System message must be at the beginnin...\n                                           ^\nError: Jinja Exception: System message must be at the beginning.","type":"server_error"}}

Anybody tried it on MLX? I was really skeptical about them releasing 27b, after they skipped it for 3.7 but I'm pleasantly surprised.

Goddamn, number go up in many benches.

So far, I’ve had 3.8 do a code review and also review some creative writing. It’s quite a bit more thorough than 3.6.