🧩 Artificial Analysis запустила AA-3D Модели получают текстовый бриф и должны самостоятельно собрать интерактивную сцену на three.js в офлайн-песочнице. В тест вошли 55 заданий в семи категориях, включая архитектуру, машины, живые организмы, игры и космос. На одну задачу модель получает до 3 часов и 250 шагов. Результаты оценивают люди вслепую: более 1000 участников отдали свыше 10 000 голосов, сравнивая пары сцен. В лидерах оказались: • Claude Sonnet 5.5 — 1194 Elo • Claude Opus 5.5 — 1190 • GPT-6 Astra — 1165 • Claude Fable 5.1 — 1131 Интересно, что качество 3D-сцен сильно коррелирует с общей «интеллектуальностью» моделей: корреляция составила 0,92. При этом стоимость одной сцены варьируется от $0,20 до $32,08. Самая дешёвая модель, GPT-6 Luna, заняла восьмое место. [Источник]()
From the original source:
Telegram — GPT Time | Нейросети | ИИ
Read full article at Telegram — GPT Time | Нейросети | ИИThis material was compiled from public sources. If its publication affects your rights → Request removal
Artificial Analysis Launches AA-3D
🧩 Artificial Analysis launches AA-3D. Models receive a text brief and must independently assemble an interactive three.js scene in an offline sandbox. The test comprised 55 tasks across seven categories, including architecture, vehicles, living organisms, games, and space. For each task, the model…
AI Summary & Key Takeaways
- 🧩 Artificial Analysis launches AA-3D
- Models receive a text brief and must independently assemble an interactive three.js scene in an offline sandbox
- The test comprised 55 tasks across seven categories, including architecture, vehicles, living organisms, games, and space
Artificial Analysis Launches AA-3D
Related stories
Co-founder of Replika launched Wabi - an AI messenger that creates applications itself
🤖 The co-founder of Replika launched Wabi - an AI messenger that creates applications itself. Evgenia Kuida opened access to Wabi - a personal AI agent that combines a messenger and a mini-application generator. Instead of the usual chats, here you can create separate threads for different...
Anthropic pitches AI safety bill that may block a model if 50+ people could be killed
The proposed policy seems to offer protections against AI run amok but actually forbids federal agencies from blocking a model's release without the permission of a judge.
Hugging Face CEO warns that Netflix’s OpenAI hack documentary isn’t accurate
Before you share the Netflix documentary "AI Gone Rogue" take note. Clem Delangue, Hugging Face's co-founder who is featured in the film, says it went "full doomer mode" and isn't accurate.
SpaceX's AI business exceeds investor expectations
SpaceX Reports AI Profit Growth, Beating Expectations.

