The Qwen 3.8 27B model has achieved a score of 52 on the Artificial Analysis Intelligence Index, matching the performance of GPT-5.6 Luna (max). This result places it just one point behind GLM-5.2 (max) and DeepSeek V4 Pro 0813 (max).
- Qwen 3.8 27B scores 52 on the Artificial Analysis Intelligence Index.
- The score is identical to GPT-5.6 Luna (max).
- It trails GLM-5.2 (max) and DeepSeek V4 Pro 0813 (max) by one point.
- GLM-5.2 has 753B parameters, while DeepSeek V4 Pro 0813 has 1.6B parameters.
The article highlights the model's strong performance relative to larger competitors.