Connect with us

Technology

Tencent Launches Hy4: Joining the Ranks of the World’s Leading Open-Source AI Models

Published

on

SHENZHEN, China, Nov. 4, 2026 /PRNewswire/ — Tencent (00700.HK) today announced the release of Tencent Hy4, its latest-generation large language model, marking a significant leap in performance and capabilities that now ranks it among the world’s leading open-source AI models. Tencent Hy plans to release model weights and system cards in the coming days, reinforcing its commitment to responsible open-source AI development, with a focus on cybersecurity and biological safety.

The new model also introduces visual understanding capabilities for the first time, alongside major improvements in instruction following, long-context comprehension and long-horizon agentic tasks, delivering stronger overall performance across a diverse set of applications.

Through advanced visual understanding, reasoning, and tool-calling capabilities, Hy4 now supports a broader range of tasks, including fine-grained image processing, complex chart analysis, and mathematical visual reasoning.

Additionally, Hy4 introduces computer-use capabilities for the first time, enabling it to interact with both command-line interfaces and graphical user interfaces to perform a wide range of work and everyday tasks. The model has undergone specialized training on professional software such as Blender and CAD, significantly enhancing its ability to plan and execute long-horizon and sophisticated tasks.

Deep Model-Product Co-Design Unlocks New Opportunities for Real-World Productivity

Leveraging significant advances in its foundational capabilities and deep model-product co-design, Hy4 unlocks a broader range of real-world productivity scenarios:

Office Productivity & Collaboration: Hy4 has significantly enhanced the user experience on Tencent WorkBuddy. Evaluation benchmarks show an average performance score of 83.18 and a pass rate of 96.15%, with Hy4 ranking first place in nine out of twelve scenarios within the mainland Chinese market. In practical use, Hy4 balances efficiency and effectiveness for visual creative tasks, 3D scene construction, and web development. Furthermore, it delivers marked improvements in document presentation, stability, and usability for human-AI collaborative tasks on WorkBuddy.

Leveraging enhanced in-context learning capabilities, Hy4 excels at understanding complex information from diverse sources. Integrated with WeCom via WorkBuddy, Hy4 connects with daily workflow data across chats, emails, documents, calendars, and meetings. When managing multiple projects simultaneously, Hy4 automatically tracks project history and recent developments. It identifies outstanding tasks, flags upcoming meetings requiring preparation, and monitors approaching deadlines, empowering users to prioritize effectively and stay aligned on important follow-ups. In addition, new team collaboration capabilities powered by Hy4 will be introduced with the upcoming release of WorkBuddy’s “Projects” feature.

Hy4 has also been optimized for professional financial analysis. By training on thousands of specialized workspaces and tens of thousands of high-quality tasks derived from real-world scenarios, the model gains a deep understanding of expert workflows in fund research, securities analysis, and quantitative investment strategies.

In a rigorous blind evaluation conducted by the Tencent Hy team in partnership with Tencent Financial Technology, Hy4 emerged as the top-performing open-source model. The assessment involved 213 financial professionals, including 153 experts from leading external institutions, approximately 70% of whom possessed more than five years of industry experience. Based on performance across 579 real-world financial tasks, Hy4 achieved the highest overall performance score.

Evaluation scores: Hy4 (3.04), DeepSeek-V4.1-Flash-1017 (2.97), Kimi K3.1 Pro (2.68), and GLM-5.3 (2.68).

In addition, Hy4 achieved a 12.03 percentage-point improvement over Hy4 Preview in the internal Investment Research Agent Benchmark (iRaB) evaluation conducted by financial AI company Rabyte.

To enhance the model’s real-world applicability across financial and other workplace scenarios, Tencent Hy has also introduced targeted search optimization that focuses on retrieving up-to-date information, integrating multiple information sources, and verifying factual consistency. In internal evaluations, Hy4 reduced its hallucination rate from 12.95% in the Hy4 Preview to 7.2%, while errors caused by unsupported assumptions fell from 23.1% to 5.6%.

Frontend & Game Development: Hy4 is optimized for high-complexity tasks, including intricate business logic, state management, front-end and back-end orchestration. This ensures that generated interfaces are not only visually polished but also fully functional. Furthermore, through collaboration with Tencent Interactive Entertainment Group (IEG) professional developers, Hy4 has significantly advanced its game development capabilities. By deeply integrating with major engines like Unreal Engine (UE), Unity, and Godot, the model now supports complex game logic and multiplayer implementations. Compared to Hy4 Preview, Hy4 improved by 19.1% in GameDev and 9.8% in GameLogic benchmarks.

Scientific Research: Hy4 combines advanced reasoning, scientific programming, and long-horizon task execution to support complex, end-to-end research workflows. Validated with TenureTrail, which assessed 123 executable tasks across seven disciplines, the model demonstrated superior performance in mathematical construction, physical modelling, environmental and earth sciences, systems engineering, and socio-economic decision-making.

Hy4 is advancing into the next stage of Recursive Self-Improvement (RSI) to further enhance its model research and development capabilities. Leveraging the Hyra 2.0 harness, Hy4 optimised the pre-training infrastructure for smaller models sharing the same architecture. Additionally, the Hy team has open-sourced two foundational models developed with Hy4 and Hyra 2.0 for single-cell representations (HyCell) and the physics fields (HyPDE). These tools are designed to provide the research community with robust, reusable assets for scientific discovery.

Advancing Responsible Open-Source AI through Cybersecurity and Biosafety Safeguards

Tencent Hy is committed to advancing responsible open-source AI, incorporating safeguards throughout model training and release, with a strong emphasis on cybersecurity and biosafety to support the safe and responsible use of AI.

Cybersecurity: Hy4 is trained on defensive security data, prioritizing proof-of-concept (PoC) development and vulnerability remediation. The model incorporates reverse engineering and penetration testing for analysis purposes. However, offensive materials related to exploitation are explicitly excluded from the training dataset.

Cybersecurity experts from Tencent’s Xuanwu Lab and Keen Security Lab conducted security assessments covering cryptographic reverse engineering, binary deobfuscation, vulnerability exploitation, and more. The assessments found no significant concerns regarding Hy4’s offensive cybersecurity capabilities in the areas evaluated.

Biosafety: Based on internal expert reviews, Hy4 demonstrates strong foundational biological knowledge; however, its capabilities in open-ended experimental design are more limited compared to frontier models.

Delivering Greater Cost Performance, Hy4 Flash Recently Released

Hy4 is currently available across Tencent Cloud TokenHub, WorkBuddy, CodeBuddy, Yuanbao, WeCom’s AI Assistant (Dayuan), ima, and more. Additionally, the model’s API is being made available via third-party platforms such as OpenRouter, Hermes, and OpenCode. Integration efforts are underway for other Tencent products and platforms, including QQ, Honor of Kings, Game of Peace, WeRead, and Tencent News.

Guided by a core commitment to practical AI, Hy4 continues to push the boundaries of AI performance while delivering exceptional cost-effectiveness. Despite significant performance improvements over Hy4 Preview, Hy4 maintains the same pricing: approximately S$0.90 per million input tokens, US$2.70 per million output tokens and US$0.045 per million cached input tokens.

Hy4 Flash is scheduled for release soon and will deliver greater cost-efficiency for users who prioritize high-value applications.

From November 4 to 15, WorkBuddy and CodeBuddy will be available for free. In addition, Hy4 will be available on Tencent Cloud TokenHub at 50% off its standard pricing, with this offer running through December 31, 2026.

About Tencent Cloud

Tencent Cloud, one of the world’s leading cloud companies, is committed to creating innovative solutions to resolve real-world issues and enabling digital transformation for smart industries. Through our extensive global infrastructure, Tencent Cloud provides businesses across the globe with stable and secure industry-leading cloud products and services, leveraging technological advancements such as cloud computing, Big Data analytics, AI, IoT, and network security. It is our constant mission to meet the needs of industries across the board, including the fields of gaming, media and entertainment, finance, healthcare, property, retail, travel, and transportation.

View original content:https://www.prnewswire.com/apac/news-releases/tencent-launches-hy4-joining-the-ranks-of-the-worlds-leading-open-source-ai-models-302929177.html

SOURCE Tencent

Continue Reading

Technology

Sprocket Security Named a Leader in GigaOm Radar for Penetration Testing as a Service

Published

on

By

One year after debuting as a Challenger and Outperformer, Sprocket earns perfect scores on four criteria as it pairs expert pentesters with a fleet of AI agents.

MADISON, Wis., Nov. 20, 2026 /PRNewswire/ — Sprocket Security has been named a Leader in the GigaOm Radar for Penetration Testing as a Service (PTaaS) v5. Sprocket placed as a Leader in the report’s Agile Optimizers scorecards and earned perfect marks on test launch efficiency, attack surface integration, rules of engagement controls, and the testing audit trail.

Sprocket Security has been named a Leader in the GigaOm Radar for Penetration Testing as a Service (PTaaS) v5.

The placement marks a clear step up from 2025, when GigaOm recognized Sprocket as a Challenger and Outperformer. In the year between, Sprocket expanded its capabilities with a fleet of AI agents that work alongside its pentesters, extending how much of the attack surface gets tested and how quickly.

From Challenger to Leader

This year’s report, authored by GigaOm analyst Chris Ray, maps four buyer profiles to four quadrants. It frames five market themes around the questions buyers should ask before signing: who performs the test, how it reaches internal networks, and what evidence an auditor will accept?

“Sprocket Security earns its Leader position by pairing AI testing agents with human validation of every finding, and wrapping both in machine-enforced scope, a customer kill switch and a checksummed audit trail, so its testing keeps pace with a changing attack surface without giving up control,” Ray shared.

Sprocket’s top scores reflect how the platform runs day to day:

Test launch efficiency (5.0): Teams scope and start testing without weeks of kickoff calls and paperwork.

Attack surface integration (5.0): Testing follows the attack surface as it changes, so new assets get tested as they appear.

Rules of engagement controls (5.0): Customers set what gets tested, when, and how, with guardrails built into the platform.

Testing audit trail (5.0): Every action is logged, giving security teams and auditors a record of what was tested and what was found.

Expert Pentesters, Now With an AI Fleet

Sprocket was founded because point-in-time pentests couldn’t keep up with how companies change. Continuous testing closed part of that gap. Sprocket’s AI agents close more of it, running alongside its pentesters on the Sprocket platform:

Apex: an agentic pentester for unauthenticated web applications, mapping each app and testing its own theories about where it might break.

Link: an agentic pentester for authenticated web applications, testing what sits behind the login.

Pinion: an agentic pentester for internal networks, falling the attack paths worth chasing.

Sprocket’s pentesters direct the work and validate what the agents find, so customers get more coverage without trading away the judgment of an experienced tester.

“A year ago GigaOm called us a Challenger. Today we are a Leader, and agentic innovation is why. We put a fleet of AI agents to work next to our pentesters for all types of testing, and the results show up in these scores,” said Casey Cammilleri, CEO and founder of Sprocket Security. “Buyers are asking who is actually testing, whether it reaches the parts of the network attackers go after, and whether the results will hold up with an auditor. Our answer is experienced pentesters with agents that extend their reach, on a platform that shows its work.”

What Customers Can Expect

Organizations that work with Sprocket Security get:

Continuous pentesting that follows the attack surface as it grows, not a once-a-year snapshot.Offensive security experts with deep experience across cloud, infrastructure, and application security.AI agents that extend testing coverage, supervised and validated by those experts.Clear rules of engagement controls and a full audit trail for every test.Remediation guidance and retesting to confirm issues are fixed.

Looking Forward

Sprocket will keep expanding its AI fleet and testing coverage. Every new agent is held to Sprocket’s Safety Framework for Autonomous Offensive Security Agents, which requires agents to stay within scope, stop at proof, log every action, and let humans step in at any point. The aim is to give buyers clear answers to the questions GigaOm says to ask: who is testing, how far the testing reaches, and what evidence it leaves behind.

Security leaders can access a complimentary copy of the GigaOm Radar for Penetration Testing as a Service v5 at https://www.sprocketsecurity.com/gigaom.

About Sprocket Security

Sprocket Security provides a technology-powered continuous penetration testing platform. Named in GigaOm and Gartner reports this year, Sprocket continuously monitors customer environments and triggers hybrid testing based on each customer’s dynamic attack surface and emerging threats. Unlike legacy penetration testing, Sprocket’s continuous approach delivers real-time insights and adaptive security, giving organizations a security posture that reflects today, not their last test.

Contact Information

Sprocket Security
Marketing Department
+1(608)260-7909
marketing@sprocketsecurity.com

View original content to download multimedia:https://www.prnewswire.com/news-releases/sprocket-security-named-a-leader-in-gigaom-radar-for-penetration-testing-as-a-service-302947219.html

SOURCE Sprocket Security

Continue Reading

Technology

Change of Date for Sivers Semiconductors’ Q3 2026 Report and Invitation to Presentation

Published

on

By

KISTA, Sweden, Nov. 20, 2026 /PRNewswire/ — Sivers Semiconductors AB (STO: SIVE) today announced that the Board of Directors has decided to move the publication of the interim report for the third quarter of 2026 from Thursday, November 26, 2026 to Monday, November 30, 2026. The report will be published at approximately 6:00 PM CET (12:00 PM ET), followed by a webcast presentation at 6:30 PM CET (12:30 PM ET).

The original date coincides with the U.S. Thanksgiving holiday, and the new date has been chosen to facilitate participation by the Company’s growing base of U.S. investors and analysts. As previously communicated, Sivers publishes its quarterly reports and hosts its webcasts after the close of trading on Nasdaq Stockholm, supporting broad participation across international markets.

The presentation will be held in English by Sivers’ management and will be followed by a Q&A session. The report will be available at www.sivers-semiconductors.com, and the webcast can be followed at:

https://events.q4inc.com/attendee/675660712

For more information, please contact:

Heine Thorsgaard

CFO, Sivers Semiconductors

Tel: +46 (0)8 703 68 00

Email: ir@sivers-semiconductors.com

About Sivers Semiconductors

We are Critical Enablers of a Greener Data Economy with Energy Efficient Photonics & Wireless Solutions. Our differentiated high-precision laser and RF beamformer technologies help our customers in key markets such as AI Data Centers, SATCOM, Defense, and Telecom solve essential performance challenges while enabling a much greener footprint. Visit us at: www.sivers-semiconductors.com.

This information was brought to you by Cision http://news.cision.com

https://news.cision.com/sivers-semiconductors/r/change-of-date-for-sivers-semiconductors–q3-2026-report-and-invitation-to-presentation,c4427582

CONTACT:

Media Contact   
Tyler Weiland 
Shelton Group
+1-972-571-7834
tweiland@sheltongroup.com

Company Contact
Heine Thorsgaard
CFO 
ir@sivers-semiconductors.com

View original content:https://www.prnewswire.com/news-releases/change-of-date-for-sivers-semiconductors-q3-2026-report-and-invitation-to-presentation-302947221.html

SOURCE Sivers Semiconductors

Continue Reading

Technology

Hoymiles showcases versatile residential solar and storage solutions at RE+ 2026

Published

on

By

LAS VEGAS, Nov. 20, 2026 /PRNewswire/ — Hoymiles showcased its latest residential solar and energy storage solutions at RE+ 2026, featuring the HiFlow Pro plug-in microinverter, HiBattery 4020 series and HMS series microinverters. Designed for a range of household energy needs, the portfolio covers emerging plug-in PV applications, expandable home energy storage and rooftop solar, highlighting Hoymiles’ continued commitment to the North American market.

HiFlow Pro brings flexibility to plug-in solar

HiFlow Pro microinverter, named to TIME’s Best Inventions 2026 list, is the first plug-in microinverter in the U.S. market to meet the UL 3700 safety standard. The HiFlow Pro 1-in-1 model connects to a single solar module and weighs approximately 1.8 kg, making installation straightforward. With a startup voltage as low as 22V and peak MPPT efficiency of 99.8%, it is designed for efficient solar generation in a compact form. Built-in Wi-Fi and Bluetooth connectivity enable real-time system monitoring through the S-Miles Home app.

HiBattery 4020 X combines solar generation and storage

The HiBattery 4020 X combines solar generation and storage in one compact system, featuring a built-in 4.02 kWh battery, four MPPTs, up to 4 kW PV input and 1,200 W output. It supports Hoymiles and third-party smart meters, with MQTT connectivity for integration with external energy management platforms. The system features 48 BMS protection functions and multiple safety measures, encased in an IP66-rated enclosure. For greater storage capacity, users can add HiBattery 4020 S expansion battery stack to scale the system up to 16.08 kWh.

HMS microinverters offer flexible rooftop configurations

Hoymiles also presented three HMS microinverter models: the HMS-500-1T-NA, HMS-1000-2T-NA and HMS-2000-4T-NA. With one, two and four PV module connections respectively, the microinverters provide flexible options for different rooftop layouts.

TÜV Rheinland recognizes HoyUltra 261A’s functional safety performance

During the exhibition, TÜV Rheinland presented Hoymiles with a functional safety test report for its HoyUltra 261A C&I BESS. The testing was conducted in accordance with IEC 60730-1:2022 Annex H and UL 60730-1:2024 Annex H, with the system receiving a “Pass” result. The report provides independent third-party testing evidence for the system’s functional safety design.

With its U.S. operations and North American technical service team, Hoymiles continues to strengthen local partnerships and deliver responsive support, while developing products and services tailored to local needs to help households build more flexible and resilient energy systems.

View original content to download multimedia:https://www.prnewswire.com/news-releases/hoymiles-showcases-versatile-residential-solar-and-storage-solutions-at-re-2026-302946740.html

SOURCE Hoymiles Power Electronic Inc. USA

Continue Reading

Trending