Friday's big release was Qwen 3.8 27B, an Apache 2 licensed 27B parameter vision-capable LLM from Alibaba's Qwen research lab. I've been looking forward to this one: 27B is an excellent size for running a model on a reasonably specced laptop, and its predecessor Qwen 3.6 27B was impressive.[reference:46]
The Overthinking Problem
Qwen's documentation describes the model as defaulting to xhigh for the reasoning effort, and the LM Studio GGUF I've been trying preserves that default. This is a hilarious default. It's absolutely not a good way to run the model, especially on consumer hardware.[reference:47]
I've been finding the results extremely entertaining. I quickly ran into problems with LM Studio's default context limit of 8,192 tokens — Qwen was using them all up thinking about even the most mundane of problems. I loaded the model with the full 262,144 maximum context length and that problem went away.[reference:48]
A Pelican Riding a Bicycle in 21 Minutes
Here's the pelican riding a bicycle SVG I got from my first attempt with that increased context length. It took 21 minutes to generate, using 22,276 reasoning tokens to produce 3,223 tokens of output.[reference:49]
This is by far the best pelican SVG I've been able to generate with a model that runs on a local machine — and this Qwen is pretty small, just a 17GB file on disk. The bicycle frame is the most convincing I've seen from any local model.[reference:50]
The Trade-Off
Qwen's reasoning_effort can be adjusted: xhigh (default) for complex tasks demanding thorough analysis, medium for balancing accuracy and speed, and low for efficient reasoning optimizing for speed and cost.[reference:51] The default xhigh setting is absolutely not a good way to run the model on consumer hardware, but for tasks that genuinely require deep reasoning, the results can be spectacular. The trick is knowing when to dial it back.