Qwen 3.8 27B is crazy.
A 27B open-weight model going head to head with Claude Opus 4.6 Max. It beats Opus on SWE-bench Pro, OSWorld, AndroidWorld and LiveCodeBench, and is surprisingly close on several of the others. You can try running it completely locally on your Mac or PC
Gemini 3.7 flash + qwen 3.8 27B is a much bigger deal than the benchmarks
Most people are comparing scores. The useful part is what happens when you give each model a different job. The Stack: → Gemini 3.7 Flash becomes the strategy layer: research, reasoning, planning,
I uploaded Qwen 3.8 27B with SGLang + DFlash 2 on 4 H100s, and I was excited to see 6K decodes per second with 48 parallel connections.
I also tried Jamidusu's Myeongri consultation using Horncheon's CLI, and it was good because there were no noticeable errors and it was fast.
GLM 5.3 absolutely cooked Qwen3.8 Max 😭
Qwen 3.8 Max returned basically a black screen And somehow Qwen cost me ~11× more for the generation GLM 5.3 is looking seriously impressive right now
You have reached the end of the archive
All of qwen38