LaunchAI ModelsAugust 18, 2026

Qwen 3.8 27B matches frontier models, runs locally

Alibaba's Qwen 3.8 27B, an Apache 2 licensed 27B vision-capable LLM, scores 52 on the Artificial Analysis Intelligence Index, matching GPT-5.6 Luna (max) and one point behind GLM-5.2 (max) and DeepSeek V4 Pro 0813 (max). It runs on consumer hardware like RTX 5090, but defaults to 'xhigh' reasoning effort causing overthinking.

How this story unfolded

12 days · 4 reports · 46 community posts · from Aug 14

  1. Aug 14
  2. Aug 16
  3. Aug 18
  4. Aug 20
  5. Aug 21
  6. Aug 22
  7. Aug 23
  8. Aug 24
  9. Aug 25
  10. Aug 26
  11. Aug 27

Daily brief

Get tomorrow's AI brief in your inbox

More stories today

Open the live feed