Skip to content
Forskning· Analysis

AI Models' Cognitive Development Uneven: High Verbal, Low Perceptual

A new study analyses the cognitive profiles of generative AI models, revealing uneven development across generations, with high scores in verbal comprehension and working memory but persistent weaknesses in perceptual reasoning.

By the Aheadline editorial team·7 juli 2026·2 min read·Source: arXiv cs.AIVerifierad signalAI-generated
AI Models' Cognitive Development Uneven: High Verbal, Low Perceptual
AI Models' Cognitive Development Uneven: High Verbal, Low Perceptual
By · Policy- & EU-reporter
Last updated

What happened?

Using a psychometric framework, researchers have evaluated the cognitive abilities of generative AI models. The study employed tasks adapted from the Wechsler Adult Intelligence Scale, alongside a newly developed Artificial Intelligence Quotient (AIQ) Benchmark. Six generations of leading multimodal models from two model families were analysed.

Key facts

Publikationsdatum20 maj 2026
Verbal förståelse/Arbetsminne>98:e percentilen
Perceptuellt resonemang<1:a percentilen
Antal modellgenerationer analyserade6
Antal modellfamiljer analyserade2

Initial evaluation of leading multimodal models using tasks adapted from the Wechsler Adult Intelligence Scale revealed a profoundly uneven cognitive architecture: near-ceiling performance in verbal comprehension and working memory (>$98^{\text{th}}$ percentile) contrasted with n

Forskare, Skribenter av studien · arXiv

Why it matters

The results indicate that contemporary AI possesses an uneven cognitive architecture. This asymmetry—featuring high performance in verbal comprehension and working memory contrasted with very low scores in perceptual reasoning—is critical for understanding current AI limitations. This further influences development towards artificial general intelligence (AGI).

Who is affected?

The study primarily affects AI researchers and developers working with generative AI and multimodal models. Companies producing or implementing these systems are also concerned, as the results highlight specific areas where current AI has clear deficiencies. Users are indirectly affected by AI service performance in various applications.

What else you should know

The research has been published as a preprint on arXiv and has not yet undergone peer review.

Frequently asked questions

Quick answers about this story

Vad har hänt?
En ny psykometrisk studie har utvärderat kognitiva förmågor hos generativa AI-modeller, inklusive ledande multimodala modeller, för att förstå deras styrkor och svagheter.
När hände det?
Studien publicerades 20 maj 2026 på arXiv.
Varför spelar det roll?
Det spelar roll eftersom det avslöjar en betydande ojämnhet i AI-modellers kognitiva arkitektur, vilket är avgörande för att förstå deras nuvarande begränsningar och för att kunna utveckla mer avancerad artificiell intelligens.
Vilka kognitiva områden undersöktes?
Studien undersökte bland annat verbal förståelse, arbetsminne och perceptuellt resonemang med hjälp av anpassade uppgifter från Wechsler Adult Intelligence Scale och ett nytt AIQ Benchmark.
Original source
arXiv cs.AI·arxiv.org

The link opens in a new window and leads to the publisher's own site.

Verifierad signal

Källan har spårats automatiskt från utgivaren via Aheadlines signalkedja.

AI-verktyg i artikeln

Topics

#Models
[ STAY UP TO DATE ]

Get similar news straight to your inbox

No affiliate linksCancel anytimeGDPR-friendly
[ Frequency ]
[ What do you want to read about? ]

You'll receive updates on 2 topics.

The reader's room

Send in a question or an addition. The newsroom reads everything before it's published and replies when relevant. No AI-generated text – just people.

Sign in to submit a comment or question.

Loading comments…
How this affects you

Read the article through your role

  • Decide whether this affects strategy over 6–12 months or is just noise.
  • Discuss with leadership: do we own the right question or does ownership need to move?
  • Ask: what risk are we taking by NOT acting on this this quarter?

Generated angle — not editorial analysis of "AI Models' Cognitive Development Uneven: High Verbal, Low Pe"