返回
原文
🚨 AI News | TestingCatalog@testingcatalog
28AI 编辑部评分，满分 100
2026-08-15 05:52· 21分钟前
AI 导读

AI/ML API 用同一任务对比测试，自托管 Qwen 3.8 27B 免费生成 three.js 沉没亚特兰蒂斯场景及 20 秒镜头漫游，Claude Opus 4.6 花费 $0.21。Qwen 构建了完整神庙与发光传送门，画面元素更丰富；Opus 则呈现柱列渐隐、布局稀疏。Qwen 权重已开源至 Hugging Face，可自行运行。

AI/ML API tested Qwen 3.8 27B (self-hosted) against Claude Opus 4.6 on the same task: Creating one self-contained three.js file rendering a sunken Atlantis, plus a scripted 20-second camera flythrough.

Both models handled the full camera path, with Qwen 3.8 27B running self-hosted for free vs $0.21 for Claude Opus 4.6.

Who's winning? 🤖

AI/ML APIQwen 3.8 27B beat Claude Opus 4.6 for $0 open-weights Qwen 3.8 27B on one rented GPU vs anthropic's Claude Opus 4.6. same prompt, no edits the test: one self-co...
Anthropic编码评测/基准

来源：🚨 AI News | TestingCatalog · x.com

导出 Markdown
返回
AI 评分 28/100在 X 查看原推
导出 Markdown
🚨 AI News | TestingCatalog · @testingcatalog · X·2026-08-15 05:52·21分钟前
AI 导读

AI/ML API 用同一任务对比测试，自托管 Qwen 3.8 27B 免费生成 three.js 沉没亚特兰蒂斯场景及 20 秒镜头漫游，Claude Opus 4.6 花费 $0.21。Qwen 构建了完整神庙与发光传送门，画面元素更丰富；Opus 则呈现柱列渐隐、布局稀疏。Qwen 权重已开源至 Hugging Face，可自行运行。

AI/ML API tested Qwen 3.8 27B (self-hosted) against Claude Opus 4.6 on the same task: Creating one self-contained three.js file rendering a sunken Atlantis, plus a scripted 20-second camera flythrough.

Both models handled the full camera path, with Qwen 3.8 27B running self-hosted for free vs $0.21 for Claude Opus 4.6.

Who's winning? 🤖

AI/ML APIQwen 3.8 27B beat Claude Opus 4.6 for $0 open-weights Qwen 3.8 27B on one rented GPU vs anthropic's Claude Opus 4.6. same prompt, no edits the test: one self-co...

来源：🚨 AI News | TestingCatalog· x.com

Anthropic编码评测/基准