Home Blog Page 814

Ascendion Named a Market Leader for Agentic Services by HFS Research

Ascendion has been named a Market Leader in HFS Research’s Horizons: Agentic Services, 2026 report.

Ascendion’s client impact – innovation velocity, capital unlock, and risk management – demonstrates the potential value available when agentic AI is correctly combined with a modern AI-native methodology and trained engineers.

BASKING RIDGE, N.J., April 10, 2026 /PRNewswire/ — Ascendion, an AI-native leader in software engineering impact, has been named a Market Leader in the HFS Horizons: Agentic Services, 2026 report, published by HFS Research.

Ascendion Named a Market Leader forAgentic Services by HFS Research
Ascendion Named a Market Leader forAgentic Services by HFS Research

HFS Research, a globally recognized authority on technology and services, evaluated 36 providers on their ability to drive impact from agentic AI by delivering meaningful innovation and business value at scale. Ascendion distinguished itself through its integrated Engineering to the Power of AI (EngineeringAI) methodology, which includes ecosystem partnerships, their AAVA™ agentic platform aligned to engineering workflows, physical AI studios for client co-creation, and a growing library of reusable agents.

“We are honored once again to be recognized by HFS for what we have built as an AI-native software engineering company,” said Paul Roehrig, Chief Strategy and Marketing Officer of Ascendion. “Thousands of AI agents in the hands of disciplined engineers are delivering meaningful outcomes for business leaders who trust us to get this right. Our Engineering to the Power of AI method is changing how businesses operate at a fundamental level. As we grow, we’re changing what’s possible for how we work and live.”

Ascendion demonstrated its ability to modernize legacy systems, accelerate the software development lifecycle (SDLC), align AI agents to standardized engineering processes, humans-at-the-wheel orchestration, and scale delivery with 100% internal platform adoption.

The recognition is grounded in outcomes that redefine enterprise AI delivery. Ascendion showcased how their engineers, working with AAVA, reduced client test effort and modernized 700,000 lines of 40-year-old code for a U.S. wealth management platform in 18 months at 45% lower cost. For a top five global bank, they delivered $300M in annual savings while improving service quality for more than 10 million customers. For a U.S. healthcare payer, they launched a platform serving one million people with zero failures, and significantly lower cost. And in one of the largest agentic AI implementations so far, Ascendion freed engineers to focus on innovation while accelerating time-to-market by 40%, and unlocked more than $500M in capital. These are not pilots. These are production outcomes, at scale, with real accountability, delivered for some of the most demanding enterprises in the world.

HFS research findings show that productivity is a top AI priority for 74% of enterprises, so the pressure to move from pilots to scaled production has never been greater. Ascendion’s platform-led approach and trained engineering workforce are meeting that demand by turning agentic AI from a promising concept into a repeatable, scalable capability.

“Ascendion stands out for its early and consistent commitment to applying AI, now combining its agentic engineering platform AAVA with product-led engineering mindset to accelerate legacy modernization, streamline software development, and deliver measurable transformation outcomes.” said David Cushman, Executive Research Leader, HFS Research.

As global enterprises accelerate from AI experimentation to transformation at scale, Ascendion continues to expand its agentic engineering capabilities by combining platform IP, specialized AI studios, and a trained global workforce to drive outcomes that build durable competitive advantage across banking, healthcare, retail, and beyond.

To download the report please visit Ascendion Named a Market Leader in HFS Horizons: Agentic Services, 2026 | Ascendion

About HFS

HFS Research is a leading research and advisory authority on enterprise transformation, serving Fortune 500 companies with fearless insights and actionable strategies. With unparalleled access to Global 2000 executives and deep expertise in AI, automation, and digital business models, HFS empowers organizations to make confident decisions that create sustainable competitive advantage. For more information, visit www.hfsresearch.com.

About Ascendion

Ascendion is an AI-native software engineering disruptor helping businesses innovate faster, smarter, and with greater impact. We partner with enterprise clients across North America, the UK, Europe, and APAC to solve complex challenges in data, experience design, software product engineering, and workforce transformation. Powered by expert engineers, thousands of AI agents, and our Engineering to the Power of AI (EngineeringAI) method, we deliver measurable outcomes that build trust, unlock value, and accelerate growth. Learn more at ascendion.com.

Engineering to the Power of AI™, AAVA™, EngineeringAI, Engineering to Elevate Life™, Enterprise PlatformsAI, Data & InsightsAI, ExperienceAI, GCCAI, OperationsAI, Platform EngineeringAI, ProductAI, and Quality EngineeringAI are trademarks or service marks of Ascendion®. AAVA™ is pending registration. Unauthorized use is strictly prohibited.

Fragile US-Iran Ceasefire Moderates Bearish Crypto Sentiment, Bybit and Block Scholes Report Finds

DUBAI, UAE, April 10, 2026 /PRNewswire/ — Bybit, the world’s second-largest cryptocurrency exchange by trading volume, has released the latest Bybit x Block Scholes Crypto Derivatives Analytics report, highlighting how a fragile US–Iran ceasefire has tempered bearish sentiment in crypto markets while leaving broader uncertainty unresolved.

Key findings:

  • Bitcoin rebounded above $70,000 following the ceasefire announcement.
  • Market activity reflects short squeeze accompanied by fresh positioning.
  • Sentiment has improved, but remains cautious amid ongoing geopolitical risks.

The report found that digital asset markets reacted swiftly to news of a conditional two-week ceasefire between the United States and Iran, particularly due to the agreement’s provision for reopening the Strait of Hormuz. This development eased immediate geopolitical concerns and helped lift Bitcoin above the $70,000 level for the first time since late March, alongside gains in other major cryptocurrencies.

Analysis of derivatives data suggests that the rally was not solely driven by the liquidation of bearish positions. While approximately $56 million in short positions were cleared on Bybit’s Bitcoin perpetual contracts, open interest increased alongside prices, indicating that traders were actively adding new exposure. At the same time, funding rates remained relatively stable, pointing to measured risk-taking rather than aggressive positioning.

Han Tan, Chief Market Analyst at Bybit, said the ceasefire has provided some support but remains fragile. “The US-Iran ceasefire has offered some relief for risk assets, including cryptos. However, relief rallies across the risk complex have been limited by lingering doubts over a meaningful end to this Middle East conflict,” Tan said. “The fragile ceasefire remains on shaky ground.”

He added that downside risks remain pronounced if tensions escalate again.

“Should this pause in the war shatter abruptly, risk assets are bound to give up much of their recent gains,” Tan said. “As long as the Strait of Hormuz remains practically shut, a temporary halt on military strikes does little to dilute inflation risks enough to alter the global monetary policy outlook.”

Options markets also reflect a cautious shift in sentiment. While the premium for downside protection has narrowed, it has not flipped decisively bullish, indicating that investors are adopting a wait-and-see approach as they assess whether the ceasefire can lead to a more durable resolution.

The report concludes that although crypto markets have shown resilience and led the initial relief rally, the sustainability of these gains will depend on further geopolitical developments and their implications for inflation and global monetary policy.

The full Bybit x Block Scholes report is available for download.

#Bybit / #CryptoArk / #BybitLearn

About Bybit

Bybit is the world’s second-largest cryptocurrency exchange by trading volume, serving a global community of over 80 million users. Founded in 2018, Bybit is redefining openness in the decentralized world by creating a simpler, open and equal ecosystem for everyone. With a strong focus on Web3, Bybit partners strategically with leading blockchain protocols to provide robust infrastructure and drive on-chain innovation. Renowned for its secure custody, diverse marketplaces, intuitive user experience, and advanced blockchain tools, Bybit bridges the gap between TradFi and DeFi, empowering builders, creators, and enthusiasts to unlock the full potential of Web3. Discover the future of decentralized finance at Bybit.com.

For more details about Bybit, please visit Bybit Press

For media inquiries, please contact: media@bybit.com

For updates, please follow: Bybit’s Communities and Social Media

Discord | Facebook | Instagram | LinkedIn | Reddit | Telegram | TikTok | X | Youtube

JCET Releases 2025 Annual Report, Posts Record-High Full-Year Revenue and Higher Profit Before Tax

SHANGHAI, April 9, 2026 /PRNewswire/ — Today, JCET Group (SSE: 600584), a leading global provider of integrated circuit (IC) back-end manufacturing and technology services, released its 2025 Annual Report. JCET reported record full-year revenue of RMB 38.87 billion in 2025, up 8.1% year-on-year. Profit before tax (PBT) totaled RMB 1.74 billion, up 5.4% year-on-year, and net profit attributable to shareholders reached RMB 1.57 billion.

In the fourth quarter of 2025, revenue was RMB 10.2 billion, up 1.4% quarter-on-quarter. Net profit attributable to shareholders for the quarter was RMB 610 million, representing a 26.6% increase from the previous quarter.

In 2025, JCET maintained a steady and progressive operating approach, delivering another record-high annual revenue and achieving its third consecutive year of revenue growth since 2023. During the reporting period, the company’s operations across both domestic and overseas sites remained stable, with capacity utilization staying at a high level. Advanced packaging capacity, including wafer-level packaging, continued to run at high load. Revenue associated with advanced packaging reached RMB 27 billion, marking a new high. JCET further increased investment in high-end advanced packaging technologies and capacity build-out, which, together with fluctuations in certain raw material prices, led to temporary cost pressures. However, as higher value-added businesses scaled and high-end capacity came on-stream, the company’s profitability improved quarter by quarter through the year. Looking ahead, JCET will continue to assess market demand and maintain forward-looking investments in frontier technologies and critical capacity build-out, further strengthening its competitiveness over the mid to long term.

From an application perspective, JCET deepened its presence in high-growth segments including high-performance computing, storage, automotive electronics, and power management. The company advanced the introduction and mass production of new products in areas such as co-packaged optics (CPO), vertical power delivery modules, and intelligent driving. In 2025, revenue from the computing electronics, industrial and medical electronics, and automotive electronics segments increased 42.6%, 40.6% and 31.7% respectively year-on-year. During the year, JCET completed production line readiness at JCET Automotive (Shanghai) Co., Ltd., supporting automotive-grade and robotics applications. Meanwhile, JCET Microelectronics (Jiangyin) Co., Ltd. accelerated its capacity ramp to support high-end intelligent application demand.

In R&D, JCET drove systematic breakthroughs in critical advanced packaging technologies while further strengthening collaborative innovation across the industry chain. Total R&D expenses for 2025 reached RMB 2.09 billion, up 21.4% year-on-year. The company will further increase R&D investment in key technology areas, including high-density multi-dimensional heterogeneous integration, advanced bonding and interconnect technologies, and glass substrates, while accelerating the translation of R&D achievements into scalable manufacturing and commercial applications. To promote an open and collaborative advanced packaging ecosystem, JCET successfully hosted the Advanced Packaging Developers Conference (APDC) in Singapore, establishing a platform for technology exchange and joint innovation to speed implementation.

Mr. Li Zheng, CEO of JCET Group, said, “Over the past year, amid a shifting demand mix and cost volatility, JCET remained focused on application-led upgrades in advanced packaging and continued to steadily improve the quality of its operations. In 2026, JCET will capitalize on growth opportunities in high-performance computing, storage, and embodied intelligence, speeding the translation of technology breakthroughs into mass-production capability and incremental revenue. We will also further raise utilization and operating efficiency in our high-end capacity to strengthen the core competitiveness that supports long-term growth.”

For more information, please refer to the JCET FY2025 Report

About JCET Group

JCET Group is a global leader in integrated circuit back-end manufacturing and technology services. We provide comprehensive turnkey solutions, including semiconductor package integration design, wafer probing, bumping, assembly, final testing, and global drop shipments.

Utilizing advanced wafer-level packaging, 2.5D/3D packaging, System-in-Package solutions, and reliable flip chip and wire bonding technologies, we support a wide range of applications, including automotive, artificial intelligence, high-performance computing, storage, communication, smart devices, industrial and medical sectors, and power and energy. With eight manufacturing facilities across China, Korea, and Singapore, we deliver efficient supply chain solutions and maintain close collaboration with our global customers.

JCET Releases 2025 Annual Report, Posts Record-High Full-Year Revenue and Higher Profit Before Tax

SHANGHAI, April 9, 2026 /PRNewswire/ — Today, JCET Group (SSE: 600584), a leading global provider of integrated circuit (IC) back-end manufacturing and technology services, released its 2025 Annual Report. JCET reported record full-year revenue of RMB 38.87 billion in 2025, up 8.1% year-on-year. Profit before tax (PBT) totaled RMB 1.74 billion, up 5.4% year-on-year, and net profit attributable to shareholders reached RMB 1.57 billion.

In the fourth quarter of 2025, revenue was RMB 10.2 billion, up 1.4% quarter-on-quarter. Net profit attributable to shareholders for the quarter was RMB 610 million, representing a 26.6% increase from the previous quarter.

In 2025, JCET maintained a steady and progressive operating approach, delivering another record-high annual revenue and achieving its third consecutive year of revenue growth since 2023. During the reporting period, the company’s operations across both domestic and overseas sites remained stable, with capacity utilization staying at a high level. Advanced packaging capacity, including wafer-level packaging, continued to run at high load. Revenue associated with advanced packaging reached RMB 27 billion, marking a new high. JCET further increased investment in high-end advanced packaging technologies and capacity build-out, which, together with fluctuations in certain raw material prices, led to temporary cost pressures. However, as higher value-added businesses scaled and high-end capacity came on-stream, the company’s profitability improved quarter by quarter through the year. Looking ahead, JCET will continue to assess market demand and maintain forward-looking investments in frontier technologies and critical capacity build-out, further strengthening its competitiveness over the mid to long term.

From an application perspective, JCET deepened its presence in high-growth segments including high-performance computing, storage, automotive electronics, and power management. The company advanced the introduction and mass production of new products in areas such as co-packaged optics (CPO), vertical power delivery modules, and intelligent driving. In 2025, revenue from the computing electronics, industrial and medical electronics, and automotive electronics segments increased 42.6%, 40.6% and 31.7% respectively year-on-year. During the year, JCET completed production line readiness at JCET Automotive (Shanghai) Co., Ltd., supporting automotive-grade and robotics applications. Meanwhile, JCET Microelectronics (Jiangyin) Co., Ltd. accelerated its capacity ramp to support high-end intelligent application demand.

In R&D, JCET drove systematic breakthroughs in critical advanced packaging technologies while further strengthening collaborative innovation across the industry chain. Total R&D expenses for 2025 reached RMB 2.09 billion, up 21.4% year-on-year. The company will further increase R&D investment in key technology areas, including high-density multi-dimensional heterogeneous integration, advanced bonding and interconnect technologies, and glass substrates, while accelerating the translation of R&D achievements into scalable manufacturing and commercial applications. To promote an open and collaborative advanced packaging ecosystem, JCET successfully hosted the Advanced Packaging Developers Conference (APDC) in Singapore, establishing a platform for technology exchange and joint innovation to speed implementation.

Mr. Li Zheng, CEO of JCET Group, said, “Over the past year, amid a shifting demand mix and cost volatility, JCET remained focused on application-led upgrades in advanced packaging and continued to steadily improve the quality of its operations. In 2026, JCET will capitalize on growth opportunities in high-performance computing, storage, and embodied intelligence, speeding the translation of technology breakthroughs into mass-production capability and incremental revenue. We will also further raise utilization and operating efficiency in our high-end capacity to strengthen the core competitiveness that supports long-term growth.”

For more information, please refer to the JCET FY2025 Report

About JCET Group

JCET Group is a global leader in integrated circuit back-end manufacturing and technology services. We provide comprehensive turnkey solutions, including semiconductor package integration design, wafer probing, bumping, assembly, final testing, and global drop shipments.

Utilizing advanced wafer-level packaging, 2.5D/3D packaging, System-in-Package solutions, and reliable flip chip and wire bonding technologies, we support a wide range of applications, including automotive, artificial intelligence, high-performance computing, storage, communication, smart devices, industrial and medical sectors, and power and energy. With eight manufacturing facilities across China, Korea, and Singapore, we deliver efficient supply chain solutions and maintain close collaboration with our global customers.

100K Creators, 4 Months, 27x Growth: BananaPro AI Studio Emerges as the All-in-One Hub for AI Image and Video Generation

SHERIDAN, Wyo., April 9, 2026 /PRNewswire/ — BananaPro AI, a leading AI image and video generation platform, today announced that its core creation hub, BananaPro AI Studio, has surpassed 100,000 registered creators within four months of its initial rollout — with monthly new sign-ups surging from 2,200 in the first month to over 60,000 in the most recent month, representing a nearly 27x increase in monthly user acquisition.

The rapid adoption underscores growing market demand for a unified, professional-grade AI creative workspace. BananaPro AI Studio seamlessly integrates both video and image generation capabilities into a single, cohesive platform, enabling developers, creators, and businesses to generate high-quality visuals at scale.

BananaPro AI continues to integrate the latest models and upgrade its core infrastructure to deliver a more intuitive creative experience.

A Unified Ecosystem of Top-Tier Models

BananaPro AI Studio centralizes AI image and AI video generation within a single interface, offering users a range of leading models. On the image generation front, the platform features the Nano Banana Series alongside models such as Midjourney, Grok, and GPT Image.

For video generation, BananaPro AI Studio supports industry-leading engines including Seedance 2.0, Veo 3.1, and Kling 3.0, enabling the creation of dynamic, high-quality video content.

Node-Based Workflow on an Infinite Canvas

Driving workflow efficiency, BananaPro AI Studio features an advanced infinite canvas powered by a visual, node-based architecture. This allows creators to leverage multiple image and video models simultaneously and design highly customized templates.

Through an intuitive drag-and-drop interface, users can connect and arrange different generation steps. Once created, these custom workflows become fully reusable assets. Content or models within any node can be dynamically replaced or updated at any time, allowing for rapid variations without disrupting the overall creative pipeline.

Industry-Specific Templates for Commercial Use

To lower the learning curve and accelerate production, BananaPro AI Studio offers a comprehensive library of ready-to-use templates. Designed for commercial applications, these customizable presets cover high-demand scenarios such as e-commerce product showcases, virtual fashion modeling, and high-fidelity jewelry rendering.

These practical starting points enable new users to grasp the platform’s full potential and begin generating professional-grade visuals from day one.

To explore the integrated model ecosystem, harness the node-based canvas, and utilize the commercial templates, try BananaPro AI Studio for free today.

About BananaPro AI

BananaPro AI is a technology company dedicated to providing AI-powered creative solutions. Specializing in AI image and video generation for e-commerce, fashion, and design professionals, BananaPro AI empowers individuals and businesses globally to enhance their productivity and unlock new creative possibilities.

Media Contact: 
Email: support@bananaproai.com
Website: https://bananaproai.com/

Binance Supports Operation Led by the UK’s NCA that Froze $12 Million in Approval Phishing Scams

ABU DHABI, UAE, April 9, 2026 /PRNewswire/ — Binance, the leading blockchain company behind the world’s largest cryptocurrency exchange by trading volume and users, today announced its participation in and support of Operation Atlantic, the UK’s National Crime Agency (NCA)-led international law enforcement operation targeting cryptocurrency and investment scammers.

Binance, along with other private industry partners and law enforcement agencies, joined Operation Atlantic, a week-long event and effort co-hosted by the NCA, US Secret Service, Ontario Provincial Police and Ontario Securities Commission. The operation focused on identifying victims who have lost, or were at risk of losing, cryptocurrency through approval phishing scams. This type of scam occurs when criminals deceive victims into granting wallet access under the guise of investment opportunities.

On-site at the NCA’s London headquarters, Binance’s Special Investigations team deployed investigative support, including live account screening and proactive scam intelligence, as part of the international law enforcement operation. The operation identified more than 20,000 victims across the UK, Canada and the United States with over $12 million in suspected criminal proceeds frozen.

The Binance Special Investigations team on the ground operated in real-time—screening and identifying thousands of victims on behalf of the operation, delivering rapid victim identification to support law enforcement outreach and intervention efforts.

In addition to victim screening, Binance provided actionable intelligence on potential bad actors to assist with asset seizure efforts and conducted proactive research that identified scam websites still actively defrauding victims at the time of the operation. No funds were seized or frozen from Binance accounts as the criminal proceeds targeted by the operation were held outside of the Binance platform.

“Operation Atlantic is a powerful example of what is possible when international agencies and private industry work side by side,” said NCA Deputy Director of Investigations Miles Bonfield in its official press release. “This intensive action has led to the safeguarding of thousands of victims in the UK and overseas, stopped criminals in their tracks and helped save others from losing their funds. “We know that fraudsters operate globally and, together with our international partners, so will the NCA to target them wherever they are based.”

“Approval phishing is one of the most damaging types of scams targeting crypto users today, and Operation Atlantic underscores how effective crime fighting is possible when private and public partners move together to stop fraud at the source,” said Binance Senior Regional Advisor for EMEA Flavio Tonon. “The inherent transparency of the blockchain makes it difficult for criminals to hide for long while exploiting victims. We are proud to have played a meaningful role in protecting thousands of potential victims.”

Binance continues to work closely with law enforcement agencies globally to further protect the blockchain ecosystem from bad actors and ensure a secure environment in the digital assets space to advance security and innovation together. In 2025, Binance supported more than 71,000 law enforcement requests and assisted in the confiscation of more than $131 million in illicit funds.

MIT Technology Review’s EmTech AI 2026: Leading in the Era of AI Integration

MIT’s flagship AI conference returns April 21–23 on the MIT campus and online, exploring how leaders can move AI from experimentation into core business infrastructure.

CAMBRIDGE, Mass., April 9, 2026 /PRNewswire/ — MIT Technology Review‘s flagship AI conference, EmTech AI 2026, will take place April 21–23 on the MIT campus, bringing together 400 senior executives, technologists, and researchers to examine one of the most important shifts in artificial intelligence: its transition from experimentation into core business infrastructure.

As organizations move beyond pilots and proofs of concept, this year’s event focuses on what MIT Technology Review calls “the Great Integration”—the challenge of embedding AI into the systems, workflows, and decision-making processes that define modern organizations.

Curated by MIT Technology Review‘s editorial team, the 2026 agenda delivers practical insight into how leading businesses are deploying AI at scale. Sessions will explore the evolving AI stack, data readiness, governance, agentic systems, and the growing role of AI in shaping strategy and operations.

The speaker lineup brings together leaders from across industry and research, including voices from Microsoft, OpenAI, Snowflake, ServiceNow, Walmart, and more, alongside MIT faculty and MIT Technology Review editors. Speakers include:

  • Donna Morris, Chief People Officer at Walmart, on integrating AI into the workforce and systems of America’s largest private employer
  • Sulman Choudhry, Head of Engineering, ChatGPT at OpenAI, on the rise of AI platforms and integrated architectures
  • Kellie Romack, Chief Digital Information Officer at ServiceNow, on building an agentic workforce across enterprise systems
  • Duncan Crabtree-Ireland, National Executive Director at SAG-AFTRA, on the future of human expression in an AI-first world
  • Ali Farhadi, recently appointed to Microsoft’s AI intelligence team, on the future of super-intelligent ecosystems and coordinated AI systems

Editorial perspectives will also play a central role, with MIT Technology Review editor in chief Mat Honan, executive editor Amy Nordrum, and senior AI editor Will Douglas Heaven leading discussions on the signals, trends, and tensions shaping AI’s next chapter.

Key themes across the agenda include:

  • Agentic AI: The shift from copilots to systems that can act, decide, and operate across workflows
  • The New AI Stack: How data, orchestration, and security must evolve as AI becomes embedded in core systems
  • AI and Trust: How discovery, credibility, and decision-making are changing in an AI-mediated world
  • Creativity and Control: The implications of AI for authorship, ownership, and human expression

In addition to the mainstage program, attendees will gain access to exclusive on-campus experiences, including MIT campus and museum tours, hands-on workshops, editorial-led roundtable discussions, and curated networking opportunities designed to foster meaningful, high-value connections.

With attendance intentionally limited to 400 participants, EmTech AI is designed to prioritize depth, access, and real dialogue between speakers and attendees.

Tickets and last-call pricing
Tickets are limited for this exclusive event. Secure your seat today with last-call pricing.

For the full agenda and to secure your spot, visit EmTech-AI.com.

Anchored by the editorial expertise of MIT Technology Review, EmTech AI’s in-person event will feature exclusive experiences, thought-provoking interviews and keynotes, and strategy-setting case studies. Attendees sit side by side with leaders and peers across all industries in interactive Q&A sessions and experience unparalleled networking opportunities, all in an intimate setting at the Media Lab on the iconic MIT campus.

The virtual experience will include an interactive event hub, featuring livestreamed sessions, videos on demand, live chat and Q&A, and additional networking opportunities.

Members of the media may obtain additional information and request press credentials by emailing press@technologyreview.com.

EmTech AI Partners
The EmTech AI Presenting Partners are HPE, a leader in essential enterprise technology; Ascendion, helping businesses innovate faster, smarter, and with greater impact with AI-powered software engineering; and GC Cybersecurity, pioneering the world’s first fully autonomous data protection and compliance platform. Additional partners and collaborators include Box, Samsung Next, Micron, MIT Professional Education, IDEO, and the MIT Museum. For additional partnership opportunities, contact Andrew Hendler at andrew.hendler@technologyreview.com.

About MIT Technology Review

MIT Technology Review is an independent media company owned by MIT. The print publication, established in 1899, was the first-ever technology magazine; today, MIT Technology Review publishes in multiple digital formats every day, including on our site, in email newsletters, and across all major social channels. We also produce a multi-award-winning bimonthly print magazine and run one of the industry’s most highly regarded events brands, EmTech.

Our goal is to become the destination for those seeking to understand how technology is shaping our world. Subscribe. Attend. Follow: LinkedIn, Reddit, Facebook, Instagram.

Media Contact:

MIT Technology Review
press@technologyreview.com

LG Reveals Next-Gen Multimodal AI ‘EXAONE 4.5’

SEOUL, South Korea, April 9, 2026 /PRNewswire/ — LG AI Research today announced the release of EXAONE 4.5, its latest multimodal AI model capable of simultaneously understanding and reasoning across both text and images.

EXAONE 4.5 is a sophisticated Vision-Language Model (VLM) that integrates a proprietary vision encoder with a Large Language Model (LLM) into a unified architecture. This latest advancement builds on the deep technical expertise LG AI Research has accumulated since December 2021, when it developed EXAONE 1.0, Korea’s first-ever multimodal AI model.

The current model serves as a strategic stepping stone for the modality expansion of “K-EXAONE,” an ongoing project dedicated to developing a proprietary AI foundation model.

Following the successful completion of the project’s second phase this August, LG AI Research plans to accelerate its modality expansion efforts upon the confirmation of the third phase. The ultimate vision is to evolve EXAONE into a form of “Physical Intelligence”—an AI capable of understanding and making judgments within the physical world, transcending the boundaries of virtual environments.

  • EXAONE 4.5 Surpasses OpenAI’s GPT-5 mini and Alibaba’s Qwen-3-VL Across 13 Visual Assessment Benchmarks

EXAONE 4.5 excels in accurately reading and reasoning through complex documents encountered in real-world industrial settings, such as contracts, technical drawings, financial statements, and scanned documents.

LG AI Research demonstrated its competitive edge by unveiling the benchmark results for EXAONE 4.5, highlighting the multimodal model’s superior performance in visual processing and reasoning.

EXAONE 4.5 achieved an average score of 77.3 across five key STEM (Science, Technology, Engineering, and Mathematics) benchmarks, outperforming major global models including OpenAI’s GPT-5-mini (73.5), Anthropic’s Claude 4.5 Sonnet (74.6), and Alibaba’s Qwen-3 235B (77.0).

The model also demonstrated superior performance across 13 comprehensive benchmarks. These include three indicators for vision general-purpose and five indicators for document comprehension and reasoning—evaluating the ability to interpret complex information within professional literature and multimodal infographics. In these categories, EXAONE 4.5 consistently surpassed the average scores of GPT-5-mini, Claude 4.5 Sonnet, and Qwen-3-VL.

Notably, EXAONE 4.5 showcased its technical edge in coding by scoring 81.4 on LiveCodeBench v6, exceeding Google’s latest model, Gemma 4 (80.0). Furthermore, on ChartQA Pro, which assesses the ability to analyze and reason through complex charts, the model recorded a world-class score of 62.2, the highest among models in its class.

An official from LG AI Research explained, “Achieving high average scores in visual assessment indicators signifies that the AI has moved beyond simply recognizing text or unstructured data. It means the model now possesses the comprehensive reasoning capabilities to grasp context and provide accurate answers to complex questions.”

EXAONE 4.5 also delivered remarkable results in terms of operational efficiency.

Despite having 33 billion parameters (33B)—roughly one-seventh the size of the “K-EXAONE” model unveiled late last year—EXAONE 4.5 achieved comparable performance in text comprehension and reasoning. This breakthrough is the result of LG AI Research’s proprietary Hybrid Attention architecture and high-speed inference technology based on multi-token prediction.

Furthermore, LG AI Research has expanded its official language support beyond Korean and English to include Spanish, German, Japanese, and Vietnamese.

  • EXAONE 4.5 Open-Weights to Drive Ecosystem Growth and Mastering Korean Cultural Context

Following the decision to release EXAONE 3.0 as an open-weight model in August 2024—a first in Korea—LG AI Research has continued its efforts to expand the AI research ecosystem.

LG AI Research has released EXAONE 4.5 available on the global open-source platform Hugging Face, permitting its use for research, academic, and educational purposes.

Meanwhile, LG has also utilized EXAONE as an educational resource to enhance the AI development skills of young talents. Earlier this month, the company hosted the ‘LG Aimers’ Hackathon, a program dedicated to nurturing young AI experts, with a focus on developing lightweight versions of the EXAONE model.

“EXAONE 4.5 represents LG AI’s successful entry into the multimodal era, where AI understands not just text, but visual information as well,” said Jinsik Lee, Head of EXAONE Lab at LG AI Research. “Starting with this model, we will expand AI’s scope of understanding to include audio, video, and the physical environment, ultimately creating AI that can make practical judgments and take action within industrial settings.”

LG AI Research is continuing its development to make EXAONE the AI that best understands Korea’s history, culture, and social context.

To this end, LG AI Research trained EXAONE using data provided by the Northeast Asian History Foundation in January and is currently discussing collaborations with other domestic institutions holding high-quality data.

“The surge in AI models capable of speaking Korean does not equate to a true understanding of cultural sensitivity,” said Myoungshin Kim, Head of the AI Safety & Trust Office at LG AI Research. “With its built-in K-AUT(Korea-Augmented Universal Taxonomy), EXAONE is evolving to provide expressive depth alongside robust reliability, setting a new standard for culturally-aware AI.”

LG AI Research

Established in December 2020, LG AI Research serves as the AI think tank for LG Group. Its mission is to enhance the Group’s AI capabilities by solving business challenges, conducting cutting-edge AI research, and establishing and implementing ethical principles for AI technology.

Official website: https://www.lgresearch.ai/

CONTACT: CHAE Ok, ok.chae@lgresearch.ai