Qwen 27B
A mid-sized open-weight large language model from Alibaba's Qwen family, often run locally.
HN commenters like it for agentic work with consistent tool calls,123 everyday non-coding tasks,4 and private local use.56 The most common caveat is slow local inference or needing a better GPU,78 and it is not as strong as frontier models.910
Why people recommend it
Agentic tool use · 3
Handles agentic workflows and tool calls well in harnesses like OpenCode and MCP setups.
“I use 27B in plan mode and 35B MoE in act mode.”
“I use this model for agentic work”
Show 1 more
“Qwen 27b works well enough in OpenCode though which is what I use.”
Model quality · 3
Several commenters call it a great model, especially when grounded.
“Qwen 27b is a great model”
“27b is a great model, if grounded.”
Show 1 more
“I have been using the qwen 27b and it is great”
Local and private · 3
Runs locally, so commenters use it for private projects and free everyday work.
“except when using local models like Qwen-27B”
“I use this 27B for everyday tasks that aren’t actually coding.”
Show 1 more
“It succesfully identified how it worked even the obfuscated classes and methods.”
Everyday tasks · 1
Used for non-coding work such as sorting emails and analysis.
“I use this 27B for everyday tasks that aren’t actually coding.”
Code analysis · 1
Worked out how a decompiled, obfuscated Android APK functioned and documented it.
“It succesfully identified how it worked even the obfuscated classes and methods.”
Distillation target · 1
A team distilled a larger model into Qwen 27B specialists with great success.
“we used to distill GLM 5.2 into Qwen 27B specialists with great success.”
Caveats and dissent
Speed and hardware · 3
Local runs can be very slow, and some commenters need a better GPU to run it reliably.
“It succesfully identified how it worked even the obfuscated classes and methods.”
“And I can't run qwen 27b reliably at the moment, I need a better gpu.”
Show 1 more
“Qwen 27b is a great model”
Behind frontier models · 2
Commenters say it is not as good as GPT, Claude, or GLM for coding.
“I use this model for agentic work”
“I use this 27B for everyday tasks that aren’t actually coding.”
Needs grounding · 2
Commenters say it needs grounding and reasonable context for good outputs.
“27b is a great model, if grounded.”
“Qwen 27b is a great model”
Tool preference · 1
It tends to use bash commands instead of OpenCode's built-in tools unless steered.
“Qwen 27b works well enough in OpenCode though which is what I use.”
Tricky distillation · 1
Building workflows to train specialists on it is tricky.
“we used to distill GLM 5.2 into Qwen 27B specialists with great success.”
Often compared with
Who recommends it in their own words
Threads where it came up
- No, Qwen 27B isn't the sweet spot 2 mentions
- Jev meets the Trolley Problem 1 mention
- Culture Becomes a Dark Forest 1 mention
- Run Qwen3.8 27B locally: real numbers from my Mac Studio 1 mention
- Qwen3.8-Max: A New Bar for Coding and Cowork 1 mention
- DeepSeek-V4-Flash Update 1 mention
Also written as Qwen-27B. Data: qwen-27b.json. Quotes are verbatim; the verdict and theme labels are written by a model from these comments, and every clause links to its sources.
All comments 12
“If little qwen 27b can reply "50/50, 100% confident"”
“except when using local models like Qwen-27B”
“And I can't run qwen 27b reliably at the moment, I need a better gpu.”
“I use 27B in plan mode and 35B MoE in act mode.”
“we used to distill GLM 5.2 into Qwen 27B specialists with great success.”
“Qwen 27b is a great model”
“27b is a great model, if grounded.”
“I use this 27B for everyday tasks that aren’t actually coding.”
“I use this model for agentic work”
“Qwen 27b works well enough in OpenCode though which is what I use.”
“It succesfully identified how it worked even the obfuscated classes and methods.”
“I have been using the qwen 27b and it is great”