ainewsblitz.com

Breaking

AI/ML API tests self-hosted Qwen3.8-27B against Claude Opus 4.6 on a Three.js Atlantis flythrough

AI/ML API tests self-hosted Qwen3.8-27B against Claude Opus 4.6 on a Three.js Atlantis flythrough
  • Foundation Models
  • Software Dev & Coding
  • Open Source

AI/ML API ran a same-prompt coding test that asked self-hosted Qwen3.8-27B and Claude Opus 4.6 to generate one self-contained Three.js file of a sunken Atlantis city, plus a scripted 20-second camera flythrough. The tester said both models completed the full camera path, and judged the open-weight 27B scene denser, at a billed inference cost of $0 versus $0.21 for Opus 4.6.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.

$20
Read this article
$29/month
Unlimited — all 27,651 articles, the full archive, and comprehension quizzes
Save 72%
$98/year
≈ $8.17/month
Unlimited, billed once a year