QNLP Quantum Natural Language Processing
A Quantum-Enhanced Interpretable and Scalable Text-based Natural Language Processing Software Package, λambeq Gen II
Quantinuum presents λambeq, Quantinuum’s next generation of quantum natural language processing (QNLP) software.
λambeq Gen II, the most recent iteration of Quantinuum’s open-source quantum natural language processing (QNLP) software, is one significant news. λambeq Gen II, which was released on May 22, 2025, enables users to translate phrases into quantum circuits that can be operated on actual quantum hardware. In addition to incorporating current advancements in language formalism and quantum hardware, this new version expands upon five years of development.
One of the main developments in Gen II is the transition from the old DisCoCat formalism to a new mathematical underpinning known as DisCoCirc. Compared to DisCoCat, DisCoCirc is made to provide a lot richer linguistic compositional structure. Scalability to entire texts rather than simply individual sentences is supported by this new formalism, which makes it possible to simulate more expressive compositionality across languages and larger text structures. In addition, DisCoCirc guarantees that the generated quantum models are canonical that is, well-defined in terms of compositional structure and learnability and adds language neutrality, which makes it flexible across linguistic domains.
These characteristics aid in resolving problems with the interpretability of findings and the trainability of quantum models that were present in previous QNLP studies. Explainable AI (XAI) features, compositional generalisation demonstrations, and evidence of enhanced quantum performance all support the release. The λambeq tool is still open source, integrates with Quantinuum’s quantum stack, and immediately transforms syntactic structures into quantum circuits using Python and VQC (variational quantum classifier) techniques.
The display of the first scalable, error-corrected, end-to-end computational chemical workflow in history is another noteworthy announcement. Because quantum computers are ideally adapted for complicated computations describing chemical events that are now unfeasible for classical supercomputers, this development is considered as a significant step towards opening up new vistas in scientific research.
An important step towards fault-tolerant quantum simulations, this work demonstrated the first practical coupling of logical qubits and quantum phase estimation (QPE) for molecular energy computations. Using Quantinuum’s InQuanto quantum chemistry platform, the demonstration was carried out on their System Model H2 quantum computer. The accomplishment demonstrates Quantinuum’s full-stack strategy by utilising the H2 hardware’s scalable QCCD design, high-fidelity operations, all-to-all connection, conditional logic, mid-circuit measurements, and real-time quantum error correction (QEC) decoding capabilities. Customers will be able to access this workflow through InQuanto, according to Quantinuum.
A collaborative venture with Al Rabban Capital to hasten the commercial adoption of quantum technology in Qatar and the Gulf is also announced. This collaboration sets the stage for Qatar to invest up to $1 billion USD over the following ten years. This alliance aims to integrate Quantinuum’s cutting-edge quantum technology, co-create regional quantum computing applications, and expand workforce.
Quantinuum’s ambition to expand commercial reach through long-term strategic alliances is strengthened by this initiative, which indicates that the US and Qatar are committed to strengthening strategic relations and bilateral investment in future-defining industries. During Qatar discussions, hybrid quantum-classical systems which mix high-performance computing, AI, and quantum computing were emphasised for improving AI across industries. Quantinuum’s H2-1 machine, which uses less power than typical supercomputers, may help AI-powered data centres meet their energy needs.
Additionally, Quantinuum declared that the Quantum Volume (QV) of their System Model H2 had surpassed 2²³ = 8,388,608. This accomplishment satisfies a pledge made in 2020 to raise their QV by ten times annually for five years. IBM created the Quantum Volume benchmark, which takes into consideration variables like qubit count, coherence times, connection, and error rates to indicate a machine’s computing capacity. According to IBM, a greater quantum volume suggests a greater capacity to investigate answers to practical issues. According to Quantinuum, QV is a non-gamable statistic that combines pertinent elements for achieving fault tolerance. Reaching this world-record QV shows and validates Quantinuum’s demonstrably superior performance. They point out that they completed this five-year commitment earlier than expected.
HBKU’s establishment of Qatar’s first quantum computing laboratory, supported by a $10 million MoD grant, is one of the other recent news items mentioned.
