Qwen 3.8 27B scored 52 on the Artificial Analysis Intelligence Index, matching GPT-5.6 Luna and trailing only GLM-5.2 and DeepSeek V4 Pro by a single point — despite those models being 753B and 1.7T parameters respectively. This highlights remarkable efficiency at a small model scale.
Background
Qwen is Alibaba's open-weight large language model family, and the Artificial Analysis Intelligence Index is a benchmark evaluating LLM capabilities. Smaller models matching or approaching much larger ones has become a key trend in 2026.
- Source
- Simon Willison
- Published
- Aug 18, 2026 at 07:58 AM
- Score
- 6.0 / 10