• 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

再帰的自己改善を測るベンチマーク「Vals RSI Index」

12/10/2026

0 Comments

 
再帰的自己改善(recursive self-improvement、RSI)は、AIが自分自身、あるいは自分の後継となるシステムを改良し、その結果「次の改良をする力」までが高まって、改良が改良を呼んでいく正のフィードバックの過程を指しています。
Vals AIが提唱した「Vals RSI Index」は、人工知能が自律的に次世代モデルの研究開発を行う能力を外部から定量的に測定するための長時間型オープンエンドベンチマークで、圧縮、言語モデル訓練、ハーネス設計、事後学習という4つの高難度タスクを通じて各モデルを評価しています。2026年10月時点の最新版ではClaude Opus 5.5が総合首位でした。
「Vals RSI Index」について、生成AIに深掘りさせましたので、ご参照ください。
なお、生成AIによる調査・分析結果は、公開された情報だけに基づくものであり、必ずしも実情を示したものではなく、誤った情報を含む可能性があることにご留意のうえ、ご参照ください。
 
Vals RSI Index
https://www.vals.ai/benchmarks/rsi_index
 
 
The Vals RSI Index: A Benchmark for Measuring Recursive Self-Improvement
Recursive self-improvement (RSI) refers to a process in which AI improves itself or its successor systems and, as a result, becomes better at making further improvements. This creates a positive feedback loop in which each improvement drives further improvements.
The Vals RSI Index, introduced by Vals AI, is a long-horizon, open-ended benchmark designed to provide an external, quantitative measure of AI’s ability to autonomously conduct research and development on next-generation models. It evaluates models through four challenging tasks: compression, language model training, harness design, and post-training. In the latest results available as of October 2026, Claude Opus 5.5 ranked first overall.
I asked generative AI to examine the Vals RSI Index in depth. Please refer to the resulting analysis.
Please note that the findings of this AI-generated research and analysis are based solely on publicly available information, may not accurately reflect actual circumstances, and may contain errors. Please review them with these limitations 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.
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

    April 2026
    March 2026
    February 2026
    January 2026
    December 2025
    November 2025
    October 2025
    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