192 likesNew · 8d old
Qwen Concept 27B trims thinking, boosts benchmark scores
What is qwen-concept-27b?
The model rewrites nine internal weight matrices without any new training data, using its own activations. It scores higher on GPQA Diamond and LiveCodeBench while needing 27-54% fewer thinking tokens.
Who it is for
Researchers and developers who want faster, more efficient large-language models should check it out.
How to try it
HF_ENDPOINT=https://models.concept.country hf download ConceptCountry/qwen-concept-27b
The story so far
Early Signal first spotted ConceptCountry/qwen-concept-27b on November 13, 2026, with 192 likes. On its busiest day it gained 32 likes. The project is 8 days old. Early Signal files it under LLMs, Coding.
Questions about qwen-concept-27b
- How do I install qwen-concept-27b?
- Run: HF_ENDPOINT=https://models.concept.country hf download ConceptCountry/qwen-concept-27b