• Home
  • Services
  • About
  • Contact
  • Blog
  • 知財活動のROICへの貢献
  • 生成AIを活用した知財戦略の策定方法
  • 生成AIとの「壁打ち」で、新たな発明を創出する方法

​
​よろず知財コンサルティングのブログ

PoetiqによるARC-AGI-2最高記録が公式に検証

7/12/2025

0 Comments

 
2025年12月5日、AIスタートアップであるPoetiq AIは、汎用人工知能(AGI)の進捗を測る最も厳格なベンチマークの一つであるARC-AGI-2において、ARC Prize財団によって公式に検証された結果が従来の最高記録を大幅に更新する54%というスコアだったと発表しました。(11月に発表されていた自社テスト結果では60%を超えていましたが、公式検証ではやはり数パーセント落ちていました。とはいえ、画期的なスコアです。)
Poetiqの成功は、既存の大規模言語モデル(LLM)の外部に「Mind Evolution」(精神の進化)メタシステムを構築し、進化アルゴリズムを用いて推論プロセスを動的に改良する、新しい推論パラダイムに基づいています。このアプローチにより、同社はGoogleのGemini 3 Deep Thinkなどの競合システムを凌駕しつつ、タスクあたりの実行コストを競合の半分以下(約30.57ドル)に削減し、性能と効率の両立を実証しました。
この内容を生成AIに調べさせました。そして、この生成AI報告内容をNotebookLMでインフォグラフィック、スライド資料にまとめさせました。
なお、生成AIによる調査・分析結果は、公開された情報からだけの分析であり、必ずしも実情を示したものではないこと、誤った情報も含まれていることについてはご留意されたうえで、ご参照ください。
 
Poetiq Shatters ARC-AGI-2 State of the Art at Half the Cost
We are proud to confirm that our system has officially outperformed existing methods, establishing a new state-of-the-art by a significant margin.
December 5, 2025
https://poetiq.ai/posts/arcagi_verified/
 
 
Poetiq’s ARC-AGI-2 Record Officially Verified
On December 5, 2025, AI startup Poetiq AI announced that the results officially verified by the ARC Prize Foundation for ARC-AGI-2—one of the most rigorous benchmarks for measuring progress toward artificial general intelligence (AGI)—showed a score of 54%, dramatically surpassing the previous record. (Although the company’s internally released test results in November exceeded 60%, the official verification showed a drop of several percentage points. Even so, it is a groundbreaking score.)
Poetiq’s breakthrough is based on a new reasoning paradigm in which a “Mind Evolution” meta-system is built outside existing large language models (LLMs). This framework dynamically improves reasoning processes using evolutionary algorithms. Through this approach, the company has demonstrated both superior performance—outperforming competitors such as Google’s Gemini 3 Deep Think—and significantly higher efficiency, reducing per-task execution costs to less than half those of competing systems (approximately $30.57).
I had a generative AI system investigate this topic, and the resulting report was compiled into infographics and slide materials using NotebookLM.
Please note that the analysis and findings produced by the generative AI are based solely on publicly available information and may not accurately represent actual circumstances; inaccuracies may also be included. Please review the content with this in mind.

Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
Your browser does not support viewing this document. Click here to download the document.
0 Comments



Leave a Reply.

    著者

    萬秀憲

    アーカイブ

    September 2025
    August 2025
    July 2025
    June 2025
    May 2025
    April 2025
    March 2025
    February 2025
    January 2025
    December 2024
    November 2024
    October 2024
    September 2024
    August 2024
    July 2024
    June 2024
    May 2024
    April 2024
    March 2024
    February 2024
    January 2024
    December 2023
    November 2023
    October 2023
    September 2023
    August 2023
    July 2023
    June 2023
    May 2023
    April 2023
    March 2023
    February 2023
    January 2023
    December 2022
    November 2022
    October 2022
    September 2022
    August 2022
    July 2022
    June 2022
    May 2022
    April 2022
    March 2022
    February 2022
    January 2022
    December 2021
    November 2021
    October 2021
    September 2021
    August 2021
    July 2021
    June 2021
    May 2021
    April 2021
    March 2021
    February 2021
    January 2021
    December 2020
    November 2020
    October 2020
    September 2020
    August 2020
    July 2020
    June 2020

    カテゴリー

    All

    RSS Feed

Copyright © よろず知財戦略コンサルティング All Rights Reserved.
サイトはWeeblyにより提供され、お名前.comにより管理されています
  • Home
  • Services
  • About
  • Contact
  • Blog
  • 知財活動のROICへの貢献
  • 生成AIを活用した知財戦略の策定方法
  • 生成AIとの「壁打ち」で、新たな発明を創出する方法