#Qwen3.8-27B

1 articles tagged Qwen3.8-27B

Tagged articles

Articles tagged Qwen3.8-27B, newest first.

Compact single-slot GPU on a desk beside a small open server case, wide monitor showing code and a terminal
AI Performance

ShapeLearn Qwen3.8-27B on a 16GB GPU: Agent Tests Pass 4 of 5 at 128k Context

ByteShape's IQ3_XS Qwen3.8-27B fits a 128k context in 16GB of VRAM with Q8 KV cache. About 31 tokens per second generation on a 70 watt RTX 2000 Ada running llama.cpp. Perfect recall to 256k tokens, 83 percent reasoning, 77 percent HumanEval Remix. Four of five agent tasks passed; the Godot level failed in thinking loops. Set an output token limit in the system prompt or long code gets truncated.

Get new articles by RSS or podcast

Add the RSS URL to your RSS reader or the podcast URL to your podcast app to get every new article as soon as it is published. The podcast reads out the headline and key points.

How to subscribe