Rohan Paul· @rohanpaul_ai · X·· 2026-09-05AI 评分54
AI 导读
AI/ML API 用四个场景简报测试 OpenAI 的 GPT‑6 Astra 与 Anthropic 的 Claude Fable 5.1，要求各自一次性生成独立 HTML 场景文件、不做修改。
正文 · 原文
GPT Astra just beat Claude Fable 5.1 on this test by
AI/ML API (@aimlapi ) .
Astra used 33% fewer tokens than Claude Fable 5.1 and cost 33% less.
GPT Astra: 25,159 tokens, $1.67
Claude Fable 5.1: 37,767 tokens, $2.51
GPT Astra and Claude Fable 5.1 were each asked to turn 4 briefs into self-contained HTML scenes. No external editing pass, one shot per scene.
This kind of setup tests several things: interpreting the scene, deciding how to represent it procedurally, writing enough code to create the geometry and motion, and getting the whole program to work on the first attempt.
GPT Astra just cooked Claude Fable 5.1 — for a lower price! OpenAI just rolled out GPT‑6 Astra — we put it against Anthropic's Fable 5.1 Both models got the same four briefs: Earth colliding with Mars, a tall ship sucked into a giant whirlpool, Godzilla leveling downtown, a mega-tsunami swallowing a city. One self-contained HTML file per scene, one shot, no edits. GPT Astra: 25,159 tokens, $1.67 Claude Fable 5.1: 37,767 tokens, $2.51 both models live on aimlapi[.]com — 1000+ models, one key.在 X 查看被引用的帖子
来源：Rohan Paul · x.com