{"version":1,"type":"story","url":"https://digestai.news/story/ibm-proves-quantum-circuits-outperform-llms-on-two-problems","json":"https://digestai.news/story/ibm-proves-quantum-circuits-outperform-llms-on-two-problems.json","markdown":"https://digestai.news/story/ibm-proves-quantum-circuits-outperform-llms-on-two-problems.md","slug":"ibm-proves-quantum-circuits-outperform-llms-on-two-problems","headline":"IBM Proves Quantum Circuits Outperform LLMs on Two Problems","summary":"On September 15, 2026, IBM Research published an article proving that shallow quantum circuits outperform large language models (LLMs) in two specific problem domains: functional problems and sampling problems. The research was conducted by Srinivasan Arunachalam, Arkopal Dutt, Hari Krovi, Rik Sengupta, and Ryan Mandelbaum. They focused on transformer-based LLMs, which are a common architecture for many prominent models like GPT, Claude, and Llama. For functional problems, they used the iterated index function as an example, showing that while transformers need substantial computational resources to solve it, shallow quantum circuits can do so with just one classical AND gate. In sampling problems, related to image generation using diffusion models, they demonstrated that even when equipped with chain-of-thought capabilities, LLMs cannot efficiently match the distribution produced by a constant-depth quantum circuit. The work highlights the theoretical separation between these two computing paradigms and suggests potential applications for quantum computing in augmenting classical AI systems.","keyPoints":["IBM Research proves quantum circuits outperform LLMs on functional and sampling problems","Quantum circuits use one classical AND gate to solve iterated index function problem","LLMs cannot efficiently sample distributions produced by shallow quantum circuits"],"whyItMatters":"This research could pave the way for concrete benchmarks comparing quantum systems and LLMs, potentially leading to new applications of quantum computing in augmenting classical AI.","category":{"slug":"research","name":"Research","url":"https://digestai.news/category/research"},"entities":{"companies":["IBM"],"models":[],"people":["Srinivasan Arunachalam","Arkopal Dutt","Hari Krovi","Rik Sengupta","Ryan Mandelbaum"]},"firstPublishedAt":"2026-09-15T15:42:01Z","updatedAt":"2026-09-15T15:42:01Z","sourceCount":1,"hasPrimarySource":false,"sources":[{"outlet":"Unite.AI","title":"IBM Research Proves Quantum Circuits Outperform LLMs on Two Problems","url":"https://unite.ai/ibm-research-proves-quantum-circuits-outperform-llms-on-two-problems","publishedAt":"2026-09-15T15:42:01Z","type":"press","primary":false,"lead":true}],"sourceNotes":null,"discussions":[],"thread":null,"cite":{"text":"Digest AI, \"IBM Proves Quantum Circuits Outperform LLMs on Two Problems\", 15 September 2026, https://digestai.news/story/ibm-proves-quantum-circuits-outperform-llms-on-two-problems","publisher":"Digest AI","title":"IBM Proves Quantum Circuits Outperform LLMs on Two Problems","datePublished":"2026-09-15T15:42:01Z","url":"https://digestai.news/story/ibm-proves-quantum-circuits-outperform-llms-on-two-problems"},"generatedBy":"Written by Digest AI's editorial model from the linked sources; the sources are the record.","license":"Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse"}