Introduction
An AI content creator puts Alibaba’s Qwen 3.8 (2.4 trillion parameters, open‑source soon) to the test, comparing it head‑to‑head with Fable 5, GPT‑5.6, and Kimi K3 across games, simulations, and UI design. The video aims to see if Qwen 3.8 lives up to its claim of being “second only to Fable 5.”
Key Comparisons & Test Results
- Racing game 🏎️ – Qwen 3.8 delivered a realistic 3D look, far more fun and polished than Fable 5’s basic output.
- Fireworks display 🎆 – GPT‑5.6 had a cleaner, more elegant UI; Qwen 3.8 lagged in front‑end quality.
- Aurora & black hole simulations 🌌 – Qwen 3.8’s black hole was neat and well‑organized; Fable 5’s version was glitchy.
- Cloth simulation 👕 – Fable 5 won for physics and interactivity; Qwen 3.8 felt less intuitive.
- Neon blaster & dragonfly games 🎮 – Qwen 3.8 excelled with 3D, smooth gameplay, and vibrant colors, outperforming both Fable 5 and GPT‑5.6.
- Flight simulator ✈️ – GPT‑5.6 worked correctly (take‑off, controls), while Qwen 3.8’s camera angle was broken.
- Remotion promo video 🎬 – Fable 5 crushed it with elegant timing and design; Qwen 3.8 looked generic.
- Dragon Realm RPG 🐉 – Kimi K3 felt open‑world with beautiful detail; Qwen 3.8’s version was basic and robotic.
- GTA‑style & Doom 🕹️ – Qwen 3.8 impressed with camera control and fun gameplay, but graphics were often buggy; Fable 5’s Doom was more playable.
- Nordic crypt game ⚔️ – Qwen 3.8 shone with great lighting and first‑person feel, beating Fable 5 and GPT‑5.6.
Overall Verdict
The creator rates Qwen 3.8 as a solid model that holds its own against frontier rivals, but it falls short of the bold “second only to Fable 5” claim. Kimi K3 remains the top choice for front‑end coding and polished games. Access is also tricky (region blocks, clunky token platform).
Final Takeaway
Qwen 3.8 is a capable step up from Qwen 3.7, but it doesn’t match Kimi K3 or Fable 5 in overall quality. For serious projects, stick with Kimi K3 or Fable 5 – Qwen 3.8 is worth trying but likely won’t blow you away.