Connect with us

Technology

Hon Hai Research Institute Launches Traditional Chinese LLM With Reasoning Capabilities

Published

1 year ago

on

First version by AI Research Center performs well in mathematics and reasoning

TAIPEI, March 10, 2025 /PRNewswire/ — Hon Hai Research Institute announced today the launch of the first Traditional Chinese Large Language Model (LLM), setting another milestone in the development of Taiwan’s AI technology with a more efficient and lower-cost model training method completed in just four weeks.

The institute, which is backed by Hon Hai Technology Group (“Foxconn”) (TWSE:2317), the world’s largest electronics manufacturer and leading technological solutions provider, said the LLM – code named FoxBrain – will be open sourced and shared publicly in the future. It was originally designed for applications used in the Group’s internal systems, covering functions such as data analysis, decision support, document collaboration, mathematics, reasoning and problem solving, and code generation.

FoxBrain not only demonstrates powerful comprehension and reasoning capabilities but is also optimized for Taiwanese users’ language style, showing excellent performance in mathematical and logical reasoning tests.

“In recent months, the deepening of reasoning capabilities and the efficient use of GPUs have gradually become the mainstream development in the field of AI. Our FoxBrain model adopted a very efficient training strategy, focusing on optimizing the training process rather than blindly accumulating computing power,” said Dr. Yung-Hui Li, Director of the Artificial Intelligence Research Center at Hon Hai Research Institute. “Through carefully designed training methods and resource optimization, we have successfully built a local AI model with powerful reasoning capabilities.”

The FoxBrain training process was powered by 120 NVIDIA H100 GPUs, scaled with NVIDIA Quantum-2 InfiniBand networking, and finished in just about four weeks. Compared with inference models recently launched in the market, the more efficient and lower-cost model training method sets a new milestone for the development of Taiwan’s AI technology.

FoxBrain is based on the Meta Llama 3.1 architecture with 70B parameters. In most categories among TMMLU+ test dataset, it outperforms Llama-3-Taiwan-70B of the same scale, particularly exceling in mathematics and logical reasoning (For TMMLU+ benchmark of FoxBrain, please refer to Fig.1). The following are the technical specifications and training strategies for FoxBrain:

Established data augmentation methods and quality assessment for 24 topic categories through proprietary technology, generating 98B tokens of high-quality pre-training data for Traditional ChineseContext window length: 128 K tokensUtilized 120 NVIDIA H100 GPUs for training, with total computational cost of 2,688 GPU daysEmployed multi-node parallel training architecture to ensure high performance and stabilityUsed a unique Adaptive Reasoning Reflection technique to train the model in autonomous reasoning

In test results, FoxBrain showed comprehensive improvements in mathematics compared to the base Meta Llama 3.1 model. It achieved significant progress in mathematical tests compared to Taiwan Llama, currently the best Traditional Chinese large model, and surpassed Meta’s current models of the same class in mathematical reasoning ability. While there is still a slight gap with DeepSeek’s distillation model, its performance is already very close to world-leading standards.

FoxBrain’s development – from data collection, cleaning and augmentation, to Continual Pre-Training, Supervised Finetuning, RLAIF, and Adaptive Reasoning Reflection – was accomplished step by step through independent research, ultimately achieving benefits approaching world-class AI models despite limited computational resources. This large language model research demonstrates that Taiwan’s technology talent can compete with international counterparts in the AI model field.

Although FoxBrain was originally designed for internal group applications, in the future, the Group will continue to collaborate with technology partners to expand FoxBrain’s applications, share its open-source information, and promote AI in manufacturing, supply chain management, and intelligent decision-making.

During model training, NVIDIA provided support through the Taipei-1 Supercomputer and technical consultation, enabling Hon Hai Research Institute to successfully complete the model pre-training with NVIDIA NeMo. FoxBrain will also become an important engine to drive the upgrade of Foxconn’s three major platforms: Smart Manufacturing. Smart EV. Smart City.

The results of FoxBrain is scheduled to be shared for the first time at a major conference during NVIDIA GTC 2025 Session Talk “From Open Source to Frontier AI: Build, Customize, and Extend Foundation Models” on March 20.

About Hon Hai Research Institute

The institute has five research centers. Each center has an average of 40 high technology R&D professionals, all of whom are focused on the research and development of new technologies, the strengthening of Foxconn’s technology and product innovation pipeline, efforts to support the Group’s transformation from “brawn” to “brains”, and the enhancement of the competitiveness of Foxconn’s “3+3” strategy.

About Foxconn here.

View original content to download multimedia:https://www.prnewswire.com/news-releases/hon-hai-research-institute-launches-traditional-chinese-llm-with-reasoning-capabilities-302396896.html

SOURCE Hon Hai Research Institute

Related Topics:
Continue Reading
Click to comment

Leave a Reply

Your email address will not be published. Required fields are marked *

Technology

VERNAL CAPITAL ACQUISITION CORP. ANNOUNCES PRICING OF $100 MILLION INITIAL PUBLIC OFFERING

Published

2 hours ago

on

May 6, 2026

By

NEW YORK, May 5, 2026 /PRNewswire/ — Vernal Capital Acquisition Corp. (NYSE: VECA) (“Vernal”) announced the pricing of its initial public offering (the “IPO”) of 10,000,000 units at $10.00 per unit. The units are expected to trade on the New York Stock Exchange (“NYSE”) under “VECAU” beginning May 6, 2026. Each unit consists of one ordinary share and one right to receive one-fourth of one ordinary share upon consummation of an initial business combination. Upon separate trading, the ordinary shares and rights are expected to be listed on NYSE under “VECA” and “VECAR,” respectively.

D. Boral Capital LLC is acting as sole book-running manager of the offering. The underwriters have a 45-day option to purchase up to 1,500,000 additional units to cover any over-allotments. The offering is expected to close on May 7, 2026, subject to customary closing conditions.

A registration statement for these securities was declared effective by the SEC on May 5, 2026. The offering is made only by means of a prospectus. Copies of the prospectus may be obtained, from D. Boral Capital LLC, 590 Madison Ave., 39th Floor, New York, New York 10022, by telephone at (212) 970-5150 or by email at dbccapitalmarkets@dboralcapital.com.

This press release shall not constitute an offer to sell or to buy, nor shall there be any sale where such offer, solicitation or sale would be unlawful prior to registration or qualification under the applicable securities laws.

About Vernal

Vernal is a blank check company formed to effect a merger, share exchange, asset acquisition, share purchase, reorganization or similar business combination with one or more businesses. Vernal’s target search will not be limited to a particular industry or geographic region.

Forward-Looking Statements

This press release contains “forward-looking statements,” including statements regarding Vernal’s IPO. These statements are subject to risks and uncertainties that could cause actual results to differ materially. No assurance can be given that the offering will be completed on the terms described, or at all. Forward-looking statements are subject to numerous conditions, beyond Vernal’s control, including those in the Risk Factors section of Vernal’s registration statement filed with the SEC. Copies are available on the SEC’s website, www.sec.gov. Vernal disclaims any obligation to release publicly updates or revisions to any forward-looking statements to reflect any change in Vernal’s expectations, except as required by law.

Contact

Binghan Yi, CFO
binghan@vernal.com
www.vernalspac.com

View original content to download multimedia:https://www.prnewswire.com/news-releases/vernal-capital-acquisition-corp-announces-pricing-of-100-million-initial-public-offering-302763378.html

SOURCE Vernal Capital Acquisition Corp.

Continue Reading

Technology

RIVANNA nominated for MedTech Scale-Up of the Year at MedTech World Awards 2026 | North America

Published

3 hours ago

on

May 5, 2026

By

Nomination places the Charlottesville-based company among growth-stage medtech leaders recognized for commercial momentum in AI-powered clinical decision support; public voting is open through May 8

CHARLOTTESVILLE, Va., May 5, 2026 /PRNewswire/ — RIVANNA®, developer of AI-powered clinical decision-support solutions, today announced that it has been nominated for MedTech Scale-Up of the Year at the MedTech World Awards 2026 | North America. Public voting is open through Friday, May 8, 2026, with category winners to be announced at the inaugural North American Awards Gala on May 11, 2026, at the Hilton West Palm Beach in Florida.

The MedTech Scale-Up of the Year category honors a growth-stage company successfully scaling revenues, partnerships, and adoption across the global medical technology ecosystem. Nominees across the program’s 22 categories were selected through a structured process led by the MedTech World Steering Committee, with category winners determined by a combination of expert evaluation and public voting from the global MedTech community.

“We have built RIVANNA on validation earned from the most rigorous technical buyers in healthcare: competitive federal awards translated into FDA-cleared products, each paired with a commercial program that meets clinicians where they work,” said Will Mauldin, PhD, Co-founder and CEO of RIVANNA. “Being nominated for MedTech Scale-Up of the Year is a meaningful affirmation of that approach and the team executing it.”

Public voting closes Friday, May 8, 2026. Members of the MedTech community are invited to support RIVANNA’s nomination at the official voting page: vote here.

The award nomination follows a year of measurable scaling for RIVANNA:

In October 2025, RIVANNA reported on being named a finalist in MedTech Innovator’s 2025 Early-Stage Grand Prize competition, selected from nearly 1,500 global applicants to represent the top 4% of medtech innovations worldwide.In December 2025, RIVANNA reported on the U.S. Food and Drug Administration’s 510(k) clearance of its Accuro® 3S Needle Guide Kit consumables, building on existing Accuro 3S device clearance.In April 2026, RIVANNA reported on peer-reviewed findings, published in 2025 in the Journal of Emergency Medicine (DOI: 10.1016/j.jemermed.2025.11.011), showing that the Accuro® XV musculoskeletal imaging system enables non-physician operators to acquire diagnostic-quality scans after just one hour of hands-on training.In May 2026, RIVANNA reported on the U.S. Food and Drug Administration’s 510(k) clearance of the Accuro® XV Diagnostic Ultrasound System for musculoskeletal imaging, authorizing commercial use across hospital and clinic settings.The company’s clinical program now spans eight sites nationwide with more than 1,500 patients enrolled.

The 2026 MedTech World Awards | North America, powered by Blue Goat Cyber, will be presented Monday, May 11, 2026, at the inaugural North American Awards Gala at the Hilton West Palm Beach, marking the first time the MedTech World Awards have been hosted in the United States.

About the MedTech Scale-Up of the Year Award
Presented by MedTech World, the MedTech Scale-Up of the Year category recognizes growth-stage medical technology companies demonstrating strong commercial momentum, expanding partnerships, and accelerating real-world adoption. The award is one of 22 categories spanning innovation, clinical excellence, regulatory strategy, investment, and leadership across the global MedTech ecosystem.

About RIVANNA
RIVANNA® is a medical technology company developing clinical decision-support solutions powered by proprietary clinical datasets, AI models, and purpose-built imaging hardware. The company’s platform automates complex anatomical analysis at the point of care, enabling faster, more confident clinical decisions while reducing variability and expanding access to advanced capabilities. The first applications target significant market opportunities in regional anesthesia and fracture care. RIVANNA has built a proven FDA regulatory track record across its Accuro® platform, with device clearances for Accuro® 3S (spinal needle guidance) and Accuro® XV (musculoskeletal imaging), a portfolio of supporting cleared consumables, and AI software modules advancing through regulatory review. The company is backed by 100+ patents and validated through clinical partnerships with leading academic medical centers. RIVANNA is headquartered in Charlottesville, Virginia, and operates an FDA-registered, ISO 13485:2016-certified manufacturing facility. Learn more at rivannamedical.com.

View original content to download multimedia:https://www.prnewswire.com/news-releases/rivanna-nominated-for-medtech-scale-up-of-the-year-at-medtech-world-awards-2026–north-america-302763342.html

SOURCE RIVANNA

Continue Reading

Technology

D2L Launch Week Highlights Latest Product Releases

Published

3 hours ago

on

May 5, 2026

By

Latest innovations are designed to save time, simplify workflows, and help drive better learning outcomes

TORONTO, May 5, 2026 /PRNewswire/ – D2L, a global leader in learning innovation, hosted its first-ever D2L Launch Week, a four-day virtual webinar series spotlighting the company’s latest product innovations across D2L Brightspace in 2026.

Throughout the week, D2L showcased a range of product releases through live demos and practical customer use cases, highlighting how institutions, school districts and organizations can help to drive engagement and improve learning outcomes. The featured updates include enhancements to D2L Lumi for idea generation, intervention suggestions, quiz creation and summarization; tools to strengthen parent and guardian outreach; and administrative capabilities designed to help large organizations delegate course and configuration management more effectively.

“We’re proud to showcase the ways D2L continues to innovate to help make learning more personalized, efficient, and scalable,” said Christian Pantel, Chief Product Officer at D2L. “From new D2L Lumi features to enhanced communication tools and more flexible distributed administration capabilities, these updates are designed to help our customers save time, improve usability, and deliver better learning experiences at scale.”

Enhancements to D2L Lumi

Among the new capabilities were several updates to D2L’s AI-native tool, D2L Lumi, designed to improve usability, transparency, and alignment across workflows, including:

D2L Lumi Ideas: Generates assignment and discussion ideas directly within Brightspace, making it easier to generate high quality content aligned to learning outcomes.D2L Lumi Insights: Gives educators access to learning intervention suggestions, designed to provide recommended next steps based on learner data.D2L Lumi Quiz: Helps educators generate questions from multiple course content topics and includes a more streamlined question-generation workflow.D2L Lumi Summary: Supports summarization from more content sources, including nested submodules, and can give educators the ability to preview and adjust source text before summarization.

Updates to Parent and Guardian Communications

D2L also introduced new parent and guardian communication enhancements to help K-12 educators strengthen engagement beyond the classroom. Teachers can now send bulk emails to all parents and guardians associated with students in their class. For individual student outreach, teachers can also email parents and guardians of a specific learner, making it easier to share timely updates on student progress and classroom activity.

Manage Distributed Administration at Scale

Distributed Administration gives organizations more flexibility to delegate administrative responsibilities across organization levels. With Distributed Administration, administrators can manage specific areas, enabling them to oversee courses while helping to reduce bottlenecks and free up time.

Learn more about the latest product releases showcased at D2L Launch Week.

About D2L   
D2L is transforming the way the world learns, helping learners achieve more than they dreamed possible. Working closely with customers all over the world, D2L is on a mission to make learning more inspiring, engaging and human. Find out how D2L helps transform lives and delivers outstanding learning outcomes in K-12, higher education and businesses.

D2L Media Contact
PR@D2L.com
X: @D2L
© 2026 D2L Corporation.

The D2L family of companies includes D2L Inc., D2L Corporation, D2L Ltd, D2L Australia Pty Ltd, D2L Europe Ltd, D2L Asia Pte Ltd, D2L India Pvt Ltd, D2L Brasil Soluções de Tecnologia para Educação Ltda and D2L Sistemas de Aprendizaje Innovadores, S. D2 R.L de C.V., and H5P Group AS.

All D2L and H5P marks are owned by the D2L group of companies. Please visit D2L.com/trademarks for a list of D2L marks. All other trademarks are the property of their respective owners.

View original content to download multimedia:https://www.prnewswire.com/news-releases/d2l-launch-week-highlights-latest-product-releases-302763345.html

SOURCE D2L

Continue Reading

Trending