Research2h ago
Research
Qwen 3.8 27B Matches GPT-5.6 Luna Score
A score of 52 on the Artificial Analysis Intelligence Index puts Qwen 3.8 27B on par with GPT-5.6 Luna. It trails GLM-5.2 and DeepSeek V4 Pro by only one point. This efficiency is striking given the massive parameter gap. Developers now have a compact model that rivals much larger proprietary systems in raw intelligence.