🏆 Claude Opus 5.5 стала самой умной моделью в рейтинге Artificial Analysis — и одновременно подешевела на 20% На максимальном уровне рассуждений модель набрала 58 баллов в Intelligence Index — это лучший результат, который когда-либо фиксировал Artificial Analysis. По ряду агентных тестов Anthropic теперь идёт наравне с GPT-6 Astra, а в шести из десяти оценок индекса Opus 5.5 показала лучший результат. Особенно заметен прогресс в работе ИИ-агентов. В Terminal-Bench 4.0 модель набрала 59,6%, прибавив 11 пунктов относительно Opus 5, а в профессиональном AA-Briefcase достигла рекордного Elo 1822. При этом Anthropic снизила стоимость API на 20% — до $4/$20 за миллион входных и выходных токенов. Ещё сильнее подешевел кэш: чтение теперь стоит $0,20 вместо $0,50 за миллион токенов. Есть и обратная сторона: Opus 5.5 рассуждает гораздо «дольше» конкурентов. На одну задачу теста она в среднем генерировала около 119 тысяч выходных токенов против 27 тысяч у GPT-6 Astra. [Источник]()
From the original source:
Telegram — GPT Time | Нейросети | ИИ
Read full article at Telegram — GPT Time | Нейросети | ИИClaude Opus 5.5 became the smartest model in the Artificial Analysis rating - and at the same time fell in price by 20%
🏆 Claude Opus 5.5 became the smartest model in the Artificial Analysis rating - and at the same time fell in price by 20%. At the maximum level of reasoning, the model scored 58 points in the Intelligence Index - this is the best result ever recorded by Artificial Analysis. In a row...
AI Summary & Key Takeaways
- 🏆 Claude Opus 5.5 became the smartest model in the Artificial Analysis rating - and at the same time fell in price by 20%
- At the maximum level of reasoning, the model scored 58 points in the Intelligence Index - this is the best result ever recorded by Artificial Analysis
- In a row...
Claude Opus 5.5 became the smartest model in the Artificial Analysis rating - and at the same time fell in price by 20%
Related stories
AI has learned to anticipate user thoughts. And it seems it will soon be able to control even the agents that control other agents...
🤯 AI has learned to anticipate user thoughts. And it appears it will soon be able to control even the agents that control other agents. A curious feature has appeared in OpenAI Codex. The neural network analyzes conversations, work results, and independently suggests what…
If your AGENTS.md has become outdated, Claude Code may notice that the documentation already diverges from the actual code, and...
If your AGENTS.md has become outdated, Claude Code may notice that the documentation already diverges from the actual code, yet continue following old instructions. Researchers deliberately created this situation and ran 80 trials. In 79 cases, Claude Code recognized...
Designer releases his working skill for building landing pages to open access; over 8,000 people bookmarked it in a single day.
Designer releases his working skill for building landing pages to open access. In just one day, more than 8,000 people bookmarked it. 🫠 The skill handles almost the entire page foundation: structure, copy, fonts, and spacing. To install it in Claude Code and immediately...
This is what future textbooks could look like
This is what future textbooks could look like: A guy used Claude Opus 5.5 and Three.js to turn a regular static physics textbook into a full-fledged interactive 3D learning environment. Now physical phenomena can be not just read and viewed in pictures, but literally…