Interoperable blockchain developer Skyline has announced that it will deploy cAP3X, a wrapped AP3X token on the Cardano blockchain. Delivered in collaboration with the Apex Fusion Foundation, the launch of cAP3X will introduce a Cardano-native asset that expands access to the Apex Fusion ecosystem.

Operating as a wrapped utility token, cAP3X will be deployed natively on Cardano to facilitate interoperability between the Cardano and Apex Fusion networks. This will enable value flow and cross-chain engagement between the two ecosystems.

Fully compatible with Cardano wallets and DeFi protocols, cAP3X will be deeply integrated into the network’s thriving decentralised finance sector. In addition to providing initial liquidity and supporting coordination with Cardano-based DEXs, an ADA/cAP3X trading pair will be established.

Following the deployment of the wrapped token in July, users will be able to transfer cAP3X from Cardano to PRIME Chain, the UTXO Layer 1 of the Apex Fusion ecosystem, via the Skyline Bridge. This enables the secure transfer of tokens and assets between blockchain ecosystems. Once bridged to PRIME, cAP3X will be converted into native AP3X, which can be used across Apex Fusion protocols.

Once cAP3X has been bridged to PRIME Chain, it will be redeemed 1:1 for native AP3X. The AP3X token on PRIME can be staked in Apex staking pools, where users can earn up to 10% APY rewards via liquid staking mechanisms already live on the network.

The initiative is part of a broader strategy to enhance access to utility tokens and enable interoperable DeFi applications between Cardano and Apex Fusion. The wrapped token structure provides a simple and transparent on-ramp for ADA holders to engage with the Apex Fusion network while also introducing AP3X as a viable utility asset within the Cardano ecosystem.

Through combining Bitcoin’s UTXO security model with Ethereum’s smart contract programmability, PRIME Chain provides a foundation for supporting cross-chain applications. It has been designed to address growing web3 fragmentation by integrating existing blockchain networks. Operating as a Layer 1 network, PRIME Chain leverages the Ouroboros consensus mechanism for secure and decentralized staking.

Following cAP3X token deployment and liquidity provisioning, bridge activation and staking will be made available on PRIME. This will enable users to participate in liquid staking and other DeFi activities on Apex Fusion’s fully interoperable Layer 1, PRIME. Prior to the wrapped token’s launch, full technical documentation and integration guides will be released for the benefit of developers and users alike.

About Skyline

Skyline is a leading developer of interoperable blockchain solutions including Skyline Bridge that links the UTXO and EVM worlds. This connects assets, applications, and communities across multiple blockchains, unlocking new possibilities for cross-chain innovation. Skyline Bridge forms the gateway to a growing range of decentralized ecosystems, bringing web3 networks closer together.

Learn more: ​​https://skylinebridge.tech/

About Apex Fusion

Apex Fusion interconnects isolated UTXO and account-based networks through cutting edge technology to unlock blockchain’s full potential by enabling effortless asset transfer, development, and growth across diverse yet unified protocols. At the core of Apex Fusion’s ecosystem lies PRIME Chain, maintained by a decentralized network of validators employing the Ouroboros proof-of-stake consensus protocol, renowned for becoming more secure as it scales.

Learn more: https://apexfusion.org/

Insomnia Labs, a leader in blockchain-powered infrastructure and financial technology, has launched CreatorFi, a platform offering stablecoin-based credit for digital creators built on Avalanche. Debuting with a $12 million credit facility from Kamui Finance and others, CreatorFi reshapes how creators access funding by integrating financial tools directly into the platforms they already use and providing them with seamless, fast, and flexible advances against their future earnings.

Insomnia Labs has built a strong reputation as a premier blockchain partner for enterprises and emerging innovators alike, having worked with major brands including Coca-Cola, L’Oréal, Under Armour, the International Cricket Council (ICC), Michael Mann, and Chubby Group. With the launch of CreatorFi, Insomnia Labs brings its tokenization expertise directly to the creator economy.

Creators—from YouTubers and other content creators to musicians and visual artists—are at the center of one of the fastest-growing industries, projected to reach $1.49 trillion by 2034. Yet, despite this economic potential, most creators remain shut out of mainstream financial services and traditional credit markets. To date, banks have been reluctant to underwrite digital rights, while legacy industry players—like record labels or royalty management firms—often impose predatory terms, demanding up to 80% ownership in exchange for upfront capital. This leaves creators with limited and ineffective alternatives when they need funding to scale production, expand marketing, or launch new projects.

CreatorFi changes that. It gives creators the ability to borrow against future earnings—like YouTube ad revenue, music royalties, or subscription income—using smart underwriting tools and blockchain-based infrastructure. With support for both fiat and stablecoin lending, the platform enables low-cost, global access to credit with real-time, AI-powered underwriting and embedded payments directly through the platforms creators already use.

“Today digital creators generate massive income through their content, yet banks still don’t know how to underwrite these modern revenue streams,” said Billy Huang, CEO of Insomnia Labs. “At Insomnia, we believe creators are building the next generation of global brands. CreatorFi gives them the financing solutions they deserve.”

CreatorFi’s early adoption highlights the clear market demand for financing solutions tailored to the creator economy. Yoola, one of the world’s largest YouTube Multi-Channel Networks (MCNs) working with 1,200+ creators globally—including Spizee, SlivkiShow, Kika Kim, Elsa Arca, RizaNova UZ, Home Animations, MultiDo Challenge, Deny Montana, and others—is among the first organizations adopting the platform.

“At Yoola, we’ve always seen YouTube channels as real, valuable assets. While traditional banks remain skeptical, we recognize the long-term value in creators’ work and future earnings. As an official YouTube partner, we have full visibility into channel analytics and receive YouTube earnings directly before disbursing them to creators. This unique position allows us to offer a safe, transparent, and secure experience for lenders, while giving YouTubers easy access to funds—without the friction of navigating traditional banking relationships. With Insomnia, we’re reshaping creator finance—unlocking growth for creators as well as new opportunities for investors,” said Arseniy Dorofey, Head of Content Partnership and Business Development at Yoola.

“The CreatorFi team is outstanding. They are using blockchain technology to bring winning value to all participants in the network: the creators, the platforms, and the financiers. Harnessing such innovative use of blockchain for our investors is exactly why we are focused on investing in Tokenized RWA. We are proud to be able to support the deployment of CreatorFi and accompany them in their future growth.” said Hadi Kabalan, Director at Kamui Finance.

In the music sector, Record, a royalty infrastructure and intelligence platform powering the financial backbone of global entertainment, is leveraging CreatorFi to offer fair, flexible financing to its enterprise clients. Through advanced data standardization, rights mapping, and embedded payment technologies, Record streamlines complex revenue flows for top artists, rightsholders, and institutional partners.

“We’re replacing outdated systems with infrastructure built for the ever-shifting landscape of IP. CreatorFi is a natural partner in our mission to modernize how creative IP is financed, tracked, and monetized,” said Travis Garrett, CEO & Founder of Record Financial.

Powering CreatorFi with Leading Blockchain Infrastructure

Insomnia Labs has raised an initial $12 million credit facility, with the ability to scale, for CreatorFi to begin offering advances, starting with a $5 million bond program for creators on Yoola’s platform with Kamui Finance. It has also attracted backing from other real-world asset investors like Intrinsic Capital, setting the stage for continued growth as it expands its credit offerings. The platform is built on Avalanche, a high-performance blockchain offering secure, transparent, and scalable infrastructure purpose-built for asset tokenization and the future of finance. CreatorFi joins the growing ecosystem of real, institutional-grade partners and platforms deploying on the Avalanche network.

“CreatorFi marks a new era where creativity becomes collateral,” said Morgan Krupetsky, Head of Institutions & Capital Markets at Ava Labs. “Insomnia Labs’ innovative platform—built on Avalanche’s high-performance infrastructure—is transforming future earnings into real, scalable capital for creators, while making it operationally viable for a new class of investors to participate.”

CreatorFi leverages the Coinbase Developer Platform to manage stablecoin liquidity and streamline the entire disbursement process securely. By precisely controlling and tracking stablecoin funds allocated for creator loans, CDP Wallets orchestrate each payout in an accurate, transparent, and auditable manner. Coinbase Offramp further enhances this efficiency by enabling creators to instantly convert their stablecoin advances into local currencies, providing seamless global access to their capital. This integrated approach simplifies and accelerates creators’ access to funds while offering clear visibility to lenders and institutional partners.

Dan Kim, Head of Strategy, CDP: “Digital creators are today’s businesses: global, always-on, and scaling fast. But the financial system hasn’t caught up. CreatorFi is flipping the script and, with Coinbase infrastructure behind the scenes, is turning views into secure, instant, and global real-world value. This is what the future of finance should look like.”

About Insomnia Labs

Insomnia Labs is an enterprise consumer company building loyalty programs and fintech solutions for large brands and institutions, including Coca-Cola, ICC, Under Armour, L’Oréal, Chubby Group, and others. The firm brings together a group of software and data experts, enabling AI-driven decision-making, monetization, and data sharing for intellectual property and product-driven businesses. For more information, please visit https://www.insomnialabs.io/ and https://creatorfi.finance.

About Yoola

Yoola is a global media and technology company specializing in content monetization and digital rights management for creators across platforms like YouTube, TikTok, and Instagram. As one of the world’s largest multi-channel networks (MCNs), Yoola enables creators to grow, manage, and commercialize their audiences at scale, providing access to advanced analytics, brand partnerships, and localized support in over 15 languages. The company empowers digital talent with the tools and infrastructure to turn content into thriving businesses. For more information, please visit https://yoola.com/

About Kamui Finance

Kamui Finance is a multi-chain institutional investor dedicated to investing in and bringing liquidity to tokenized Real World Assets. It is backed by veteran investors from Goldman Sachs Asset Management, Bessemer Trust, and Blakeney Management, and blockchain pioneers and technologists from Emurgo, CTIA, Mitsubishi Corporation, and Syla Technologies. For more information, please visit https://kamui.finance/

Discover how CLT leverages Ethereum blockchain to solve unfair reward systems in gaming, empowering players, creators, and communities across the global e-sports ecosystem.

CLT (Chicago Coin) is an ERC-20 token created to reshape the e-sports and gaming landscape. Built on Ethereum, CLT serves as a medium of exchange, reward currency, and governance asset—empowering gamers, developers, creators, and fans through a Decentralized Autonomous Organization (DAO) model. Unlike traditional in-game currencies, CLT enables true asset ownership, cross-game compatibility, and real-world value through secure, intermediary-free smart contracts.

Token Overview

• Token name: Chicago coin
• Token symbol: CLT
• Total supply: 1,000,000,000

What are the utilities of $CLT?

CLT powers a wide range of applications across the gaming and e-sports ecosystem:

• In-Game Transactions: Purchase items, enter tournaments, and unlock premium content.
• Cross-Platform Asset Trading: Trade assets seamlessly across games via decentralized marketplaces.
• Staking: Stake CLT to earn rewards while supporting network stability.
• Governance: Vote on ecosystem upgrades, DAO policies, and reward allocations.
• Performance Rewards: Earn CLT based on gameplay achievements, content creation, or active community participation.
• Fan Engagement: Sponsor teams, vote in competitions, and receive rewards based on team performance.
• Streaming Monetization: Creators can earn transparent payouts for their content via smart contracts.

This utility-rich structure allows CLT to support both everyday gamers and elite players, fostering a merit-based economy where rewards are earned by value creation, not gatekeeping.

What Does the CLT Ecosystem Include?

The CLT ecosystem is designed to connect and reward every stakeholder in the e-sports world:

• Gamers & Streamers: Earn and spend CLT for performance, engagement, or content creation.
• Developers: Integrate CLT for in-game payments, rewards, and NFTs.
• Content Creators: Monetize tutorials, live streams, and guides through blockchain-secured smart contracts.
• Community Builders: Get rewarded for hosting tournaments, moderating, or mentoring.
• Fans: Participate actively by sponsoring teams, voting in competitions, or trading exclusive digital rights.

This circular economy model ensures that value generated by the community stays within the ecosystem, driving continuous utility, adoption, and token demand.

About Coinstore

Accessibility. Security. Equity.
As a leading global platform for cryptocurrency and blockchain technology, Coinstore seeks to build an ecosystem that grants everyone access to digital assets and blockchain technology. With over 10 million users worldwide, Coinstore aims to become the preferred cryptocurrency trading platform and digital service provider worldwide.

New Offering Delivers a Unique Fusion Architecture That’s Being Leveraged by Industry-Leading AI Pioneers Like Cohere, CoreWeave, and NVIDIA to Deliver Breakthrough Performance Gains and Reduce Infrastructure Requirements For Massive AI Training and Inference Workloads

From RAISE SUMMIT 2025: WEKA unveiled NeuralMesh Axon, a breakthrough storage system that leverages an innovative fusion architecture designed to address the fundamental challenges of running exascale AI applications and workloads. NeuralMesh Axon seamlessly fuses with GPU servers and AI factories to streamline deployments, reduce costs, and significantly enhance AI workload responsiveness and performance, transforming underutilized GPU resources into a unified, high-performance infrastructure layer.

Building on the company’s recently announced NeuralMesh storage system, the new offering enhances its containerized microservices architecture with powerful embedded functionality, enabling AI pioneers, AI cloud and neocloud service providers to accelerate AI model development at extreme scale, particularly when combined with NVIDIA AI Enterprise software stacks for advanced model training and inference optimization. NeuralMesh Axon also supports real-time reasoning, with significantly improved time-to-first-token and overall token throughput, enabling customers to bring innovations to market faster.

AI Infrastructure Obstacles Compound at Exascale

Performance is make-or-break for large language model (LLM) training and inference workloads, especially when running at extreme scale. Organizations that run massive AI workloads on traditional storage architectures, which rely on replication-heavy approaches, waste NVMe capacity, face significant inefficiencies, and struggle with unpredictable performance and resource allocation.

The reason? Traditional architectures weren’t designed to process and store massive volumes of data in real-time. They create latency and bottlenecks in data pipelines and AI workflows that can cripple exascale AI deployments. Underutilized GPU servers and outdated data architectures turn premium hardware into idle capital, resulting in costly downtime for training workloads. Inference workloads struggle with memory-bound barriers, including key-value (KV) caches and hot data, resulting in reduced throughput and increased infrastructure strain. Limited KV cache offload capacity creates data access bottlenecks and complicates resource allocation for incoming prompts, directly impacting operational expenses and time-to-insight. Many organizations are transitioning to NVIDIA accelerated compute servers, paired with NVIDIA AI Enterprise software, to address these challenges. However, without modern storage integration, they still encounter significant limitations in pipeline efficiency and overall GPU utilization.

Built For The World’s Largest and Most Demanding Accelerated Compute Environments

To address these challenges, NeuralMesh Axon’s high-performance, resilient storage fabric fuses directly into accelerated compute servers by leveraging local NVMe, spare CPU cores, and its existing network infrastructure. This unified, software-defined compute and storage layer delivers consistent microsecond latency for both local and remote workloads—outpacing traditional local protocols like NFS.

Additionally, when leveraging WEKA’s Augmented Memory Grid capability, it can provide near-memory speeds for KV cache loads at massive scale. Unlike replication-heavy approaches that squander aggregate capacity and collapse under failures, NeuralMesh Axon’s unique erasure coding design tolerates up to four simultaneous node losses, sustains full throughput during rebuilds, and enables predefined resource allocation across the existing NVMe, CPU cores, and networking resources—transforming isolated disks into a memory-like storage pool at exascale and beyond while providing consistent low latency access to all addressable data.

Cloud service providers and AI innovators operating at exascale require infrastructure solutions that can match the exponential growth in model complexity and dataset sizes. NeuralMesh Axon is specifically designed for organizations operating at the forefront of AI innovation that require immediate, extreme-scale performance rather than gradual scaling over time. This includes AI cloud providers and neoclouds building AI services, regional AI factories, major cloud providers developing AI solutions for enterprise customers, and large enterprise organizations deploying the most demanding AI inference and training solutions that must agilely scale and optimize their AI infrastructure investments to support rapid innovation cycles.

Delivering Game-Changing Performance for Accelerated AI Innovation

Early adopters, including Cohere, the industry’s leading security-first enterprise AI company, are already seeing transformational results.

Cohere is among WEKA’s first customers to deploy NeuralMesh Axon to power its AI model training and inference workloads. Faced with high innovation costs, data transfer bottlenecks, and underutilized GPUs, Cohere first deployed NeuralMesh Axon in the public cloud to unify its AI stack and streamline operations.

“For AI model builders, speed, GPU optimization, and cost-efficiency are mission-critical. That means using less hardware, generating more tokens, and running more models—without waiting on capacity or migrating data,” said Autumn Moulder, vice president of engineering at Cohere. “Embedding WEKA’s NeuralMesh Axon into our GPU servers enabled us to maximize utilization and accelerate every step of our AI pipelines. The performance gains have been game-changing: Inference deployments that used to take five minutes can occur in 15 seconds, with 10 times faster checkpointing. Our team can now iterate on and bring revolutionary new AI models, like North, to market with unprecedented speed.”

To improve training and help develop North, Cohere’s secure AI agents platform, the company is deploying WEKA’s NeuralMesh Axon on CoreWeave Cloud, creating a robust foundation to support real-time reasoning and deliver exceptional experiences for Cohere’s end customers.

“We’re entering an era where AI advancement transcends raw compute alone—it’s unleashed by intelligent infrastructure design. CoreWeave is redefining what’s possible for AI pioneers by eliminating the complexities that constrain AI at scale,” said Peter Salanki, CTO and co-founder at CoreWeave. “With WEKA’s NeuralMesh Axon seamlessly integrated into CoreWeave’s AI cloud infrastructure, we’re bringing processing power directly to data, achieving microsecond latencies that reduce I/O wait time and deliver more than 30 GB/s read, 12 GB/s write, and 1 million IOPS to an individual GPU server. This breakthrough approach increases GPU utilization and empowers Cohere with the performance foundation they need to shatter inference speed barriers and deliver advanced AI solutions to their customers.”

“AI factories are defining the future of AI infrastructure built on NVIDIA accelerated compute and our ecosystem of NVIDIA Cloud Partners,” said Marc Hamilton, vice president of solutions architecture and engineering at NVIDIA. “By optimizing inference at scale and embedding ultra-low latency NVMe storage close to the GPUs, organizations can unlock more bandwidth and extend the available on-GPU memory for any capacity. Partner solutions like WEKA’s NeuralMesh Axon deployed with CoreWeave provide a critical foundation for accelerated inferencing while enabling next-generation AI services with exceptional performance and cost efficiency.”

The Benefits of Fusing Storage and Compute For AI Innovation

NeuralMesh Axon delivers immediate, measurable improvements for AI builders and cloud service providers operating at exascale, including:

• Expanded Memory With Accelerated Token Throughput: Provides tight integration with WEKA’s Augmented Memory Grid technology, extending GPU memory by leveraging it as a token warehouse. This has delivered a 20x improvement in time to first token performance across multiple customer deployments, enabling larger context windows and significantly improved token processing efficiency for inference-intensive workloads. Furthermore, NeuralMesh Axon enables customers to dynamically adjust compute and storage resources and seamlessly supports just-in-time training and just-in-time inference.
• Huge GPU Acceleration and Efficiency Gains: Customers are achieving dramatic performance and GPU utilization improvements with NeuralMesh Axon, with AI model training workloads exceeding 90%—a three-fold improvement over the industry average. NeuralMesh Axon also reduces the required rack space, power, and cooling requirements in on-premises data centers, helping to lower infrastructure costs and complexity by leveraging existing server resources.
• Immediate Scale for Massive AI Workflows: Designed for AI innovators who need immediate extreme scale, rather than to grow over time. NeuralMesh Axon’s containerized microservices architecture and cloud-native design enable organizations to scale storage performance and capacity independently while maintaining consistent performance characteristics across hybrid and multicloud environments.
• Enables Teams to Focus on Building AI, Not Infrastructure: Runs seamlessly across hybrid and cloud environments, integrating with existing Kubernetes and container environments to eliminate the need for external storage infrastructure and reduce complexity.

“The infrastructure challenges of exascale AI are unlike anything the industry has faced before. At WEKA, we’re seeing organizations struggle with low GPU utilization during training and GPU overload during inference, while AI costs spiral into millions per model and agent,” said Ajay Singh, chief product officer at WEKA. “That’s why we engineered NeuralMesh Axon, born from our deep focus on optimizing every layer of AI infrastructure from the GPU up. Now, AI-first organizations can achieve the performance and cost efficiency required for competitive AI innovation when running at exascale and beyond.”

Availability

NeuralMesh Axon is currently available in limited release for large-scale enterprise AI and neocloud customers, with general availability scheduled for fall 2025. For more information, visit:

Product Page: https://www.weka.io/product/neuralmesh-axon/
Solution Brief: https://www.weka.io/resources/solution-brief/weka-neuralmesh-axon-solution-brief
Blog Post: https://www.weka.io/blog/ai-ml/neuralmesh-axon-reinvents-ai-infrastructure-economics-for-the-largest-workloads/

About WEKA

WEKA is transforming how organizations build, run, and scale AI workflows through NeuralMesh, its intelligent, adaptive mesh storage system. Unlike traditional data infrastructure, which becomes more fragile as AI environments expand, NeuralMesh becomes faster, stronger, and more efficient as it scales, growing with your AI environment to provide a flexible foundation for enterprise and agentic AI innovation. Trusted by 30% of the Fortune 50 and the world’s leading neoclouds and AI innovators, NeuralMesh maximizes GPU utilization, accelerates time to first token, and lowers the cost of AI innovation. Learn more at www.weka.io

In September 2024, Skywork first open-sourced the Skywork-Reward series models and related datasets. Over the past nine months, these models and data have been widely adopted by the open-source community for research and practice, with over 750,000 cumulative downloads on the HuggingFace platform, helping multiple frontier models achieve excellent results in authoritative evaluations such as RewardBench.

On July 4, 2025, Skywork continues to open-source the second-generation reward models – the Skywork-Reward-V2 series, comprising 8 reward models based on different base models of varying sizes, with parameters ranging from 600 million to 8 billion. These models have achieved top rankings across seven major mainstream reward model evaluation benchmarks.

Skywork-Reward-V2 Download Links
HuggingFace: https://huggingface.co/collections/Skywork/skywork-reward-v2-685cc86ce5d9c9e4be500c84
GitHub: https://github.com/SkyworkAI/Skywork-Reward-V2
Technical Report: https://arxiv.org/abs/2507.01352

Reward models play a crucial role in the Reinforcement Learning from Human Feedback (RLHF) process. In developing this new generation of reward models, we constructed a hybrid dataset called Skywork-SynPref-40M, containing a total of 40 million preference pairs.

To achieve large-scale, efficient data screening and filtering, Skywork specially designed a two-stage human-machine collaborative process that combines high-quality human annotation with the scalable processing capabilities of models. In this process, humans provide rigorously verified high-quality annotations, while Large Language Models (LLMs) automatically organize and expand based on human guidance.

Based on the above high-quality hybrid preference data, we developed the Skywork-Reward-V2 series, which demonstrates broad applicability and excellent performance across multiple capability dimensions, including general alignment with human preferences, objective correctness, safety, resistance to style bias, and best-of-N scaling capability. Experimental validation shows that this series of models achieved the best performance on seven mainstream reward model evaluation benchmarks.

01 Skywork-SynPref-40M: Human-Machine Collaboration for Million-Scale Human Preference Data Screening

Even the most advanced current open-source reward models still perform inadequately on most mainstream evaluation benchmarks. They fail to effectively capture the subtle and complex characteristics of human preferences, particularly when facing multi-dimensional, multi-level feedback.

Additionally, many reward models tend to excel on specific benchmark tasks but struggle to transfer to new tasks or scenarios, exhibiting obvious “overfitting” phenomena. Although existing research has attempted to improve performance through optimizing objective functions, improving model architectures, and recently emerging Generative Reward Models, the overall effectiveness remains quite limited.

We believe that the current fragility of reward models mainly stems from the limitations of existing preference datasets, which often have limited coverage, mechanical label generation methods, or lack rigorous quality control.

Therefore, in developing the new generation of reward models, we not only continued the first generation’s experience in data optimization but also introduced more diverse and larger-scale real human preference data, striving to improve data scale while maintaining data quality.

Consequently, Skywork proposes Skywork-SynPref-40M – the largest preference hybrid dataset to date, containing a total of 40 million preference sample pairs. Its core innovation lies in a “human-machine collaboration, two-stage iteration” data selection pipeline.

Stage 1: Human-Guided Small-Scale High-Quality Preference Construction

The team first constructed an unverified initial preference pool and used Large Language Models (LLMs) to generate preference-related auxiliary attributes such as task type, objectivity, and controversy. Based on this, human annotators followed a strict verification protocol and used external tools and advanced LLMs to conduct detailed reviews of partial data, ultimately constructing a small-scale but high-quality “gold standard” dataset as the basis for subsequent data generation and model evaluation.

Subsequently, we used preference labels from the gold standard data as guidance, combined with LLM large-scale generation of high-quality “silver standard” data, thus achieving data volume expansion. The team also conducted multiple rounds of iterative optimization: in each round, training reward models and identifying model weaknesses based on their performance on gold standard data; then retrieving similar samples and using multi-model consensus mechanisms for automatic annotation to further expand and enhance silver standard data. This human-machine collaborative closed-loop process continues iteratively, effectively improving the reward model’s understanding and discrimination of preferences.

Stage 2: Fully Automated Large-Scale Preference Data Expansion

After obtaining preliminary high-quality models, the second stage turns to automated large-scale data expansion. This stage no longer relies on manual review but uses trained reward models to perform consistency filtering:

• If a sample’s label is inconsistent with the current optimal model’s prediction, or if the model’s confidence is low, LLMs are called to automatically re-annotate;
• If the sample label is consistent with the “gold model” (i.e., a model trained only on human data) prediction and receives support from the current model or LLM, it can directly pass screening.

Through this mechanism, the team successfully screened 26 million selected data points from the original 40 million samples, achieving a good balance between preference data scale and quality while greatly reducing the human annotation burden.

02 Skywork-Reward-V2: Matching Large Model Performance with Small Model Size

Compared to the previous generation Skywork-Reward, Skywork newly released Skywork-Reward-V2 series provides 8 reward models trained based on Qwen3 and LLaMA3 series models, with parameter scales covering from 600 million to 8 billion.

On seven mainstream reward model evaluation benchmarks including Reward Bench v1/v2, PPE Preference & Correctness, RMB, RM-Bench, and JudgeBench, the Skywork-Reward-V2 series comprehensively achieved current state-of-the-art (SOTA) levels.

Compensating for Model Scale Limitations with Data Quality and Richness

Even the smallest model, Skywork-Reward-V2-Qwen3-0.6B, achieves overall performance nearly matching the previous generation’s strongest model, Skywork-Reward-Gemma-2-27B-v0.2, on average. The largest scale model, Skywork-Reward-V2-Llama-3.1-8B, achieved comprehensive superiority across all mainstream benchmark tests, becoming the currently best-performing open-source reward model overall.

Broad Coverage of Multi-Dimensional Human Preference Capabilities

Additionally, Skywork-Reward-V2 achieved leading results in multiple advanced capability evaluations, including Best-of-N (BoN) tasks, bias resistance capability testing (RM-Bench), complex instruction understanding, and truthfulness judgment (RewardBench v2), demonstrating excellent generalization ability and practicality.

Highly Scalable Data Screening Process Significantly Improves Reward Model Performance

Beyond excellent performance in evaluations, the team also found that in the “human-machine collaboration, two-stage iteration” data construction process, preference data that underwent careful screening and filtering could continuously and effectively improve reward models’ overall performance through multiple iterative training rounds, especially showing remarkable performance in the second stage’s fully automated data expansion.

In contrast, blindly expanding raw data not only fails to improve initial performance but may introduce noise and negative effects. To further validate the critical role of data quality, we conducted experiments on a subset of 16 million data points from an early version. Results showed that training an 8B-scale model using only 1.8% (about 290,000) of the high-quality data already exceeded the performance of current 70B-level SOTA reward models. This result again confirms that the Skywork-SynPref dataset not only leads in scale but also has significant advantages in data quality.

03 Welcoming a New Milestone for Open-Source Reward Models: Helping Build Future AI Infrastructure

In this research work on the second-generation reward model Skywork-Reward-V2, the team proposed Skywork-SynPref-40M, a hybrid dataset containing 40 million preference pairs (with 26 million carefully screened pairs), and Skywork-Reward-V2, a series of eight reward models with state-of-the-art performance designed for broad task applicability.

We believe this research work and the continued iteration of reward models will help advance the development of open-source reward models and more broadly promote progress in Reinforcement Learning from Human Feedback (RLHF) research. This represents an important step forward for the field and can further accelerate the prosperity of the open-source community.

The Skywork-Reward-V2 series models focus on research into scaling preference data. In the future, the team’s research scope will gradually expand to other areas that have not been fully explored, such as alternative training techniques and modeling objectives.

Meanwhile, considering recent development trends in the field – reward models and reward shaping mechanisms have become core components in today’s large-scale language model training pipelines, applicable not only to RLHF based on human preference learning and behavior guidance, but also to RLVR including mathematics, programming, or general reasoning tasks, as well as agent-based learning scenarios.

Therefore, we envision that reward models, or more broadly, unified reward systems, are poised to form the core of AI infrastructure in the future. They will no longer merely serve as evaluators of behavior or correctness, but will become the “compass” for intelligent systems navigating complex environments, helping them align with human values and continuously evolve toward more meaningful goals.

Additionally, Skywork released the world’s first deep research AI workspace agents in May, which you can experience by visiting: skywork.ai

Tokenizing real estate to Unveil global opportunities & fractional ownership for all investors.

STONEFORM is reshaping the real estate investment landscape by leveraging blockchain technology to create a decentralized platform for fractional property ownership, expanding global access and liquidity for investors. Through the power of tokenization, STONEFORM is set to make property ownership more accessible, efficient, and transparent by allowing fractional ownership of real estate assets.

STONEFORM’s Vision: A New Digital Paradigm for Real Estate Investment

STONEFORM’s goal is to integrate blockchain technology and decentralized finance (DeFi) to unveil the power of real estate investment. STONEFORM enables global participation, providing diverse investment options for individuals and institutions. Token holders can engage in real estate investments without the burdens typically associated with traditional property ownership.

“At STONEFORM,we are building more than just a platform; we are building a milestone in real estate, We believe blockchain is the key to facilitating widespread access to high-quality real estate assets, enabling anyone, regardless of their financial background, to invest in and benefit from the growth of this sector.” Ukrit Thaweerat, Founder.

Main Functionalities of STONEFORM

• Fractional Ownership: Purchase fractional shares of premium real estate, lowering entry barriers for small investors globally.
• Blockchain-Powered Liquidity: Tokenized assets trade on decentralized markets,ensuring faster and more cost-effective transactions.
• Smart Contracts for Automated Management: Automates property management tasks like rent distribution, reducing costs and administrative efforts.
• Decentralized Governance: Token holders vote on decisions,giving the community control over the platform’s governance and direction.
• Global Access: Blockchain enables worldwide participation in real estate investment.
• Security and Compliance: Robust security features and automated compliance checks ensure safe and regulated transactions.

A New Era for Real Estate Investment

The global real estate market is valued at trillions, but traditional investments often require large capital and have limited liquidity. STONEFORM solves these issues with blockchain-powered fractional ownership.

Conclusion

STONEFORM is redefining the way people invest in real estate by integrating blockchain and decentralized finance. With fractional ownership, smart contracts, and decentralized governance, the platform is set to make real estate investment more accessible,liquid,and transparent than ever before. The project will continue to expand its offerings, driving the future of real estate investment on a global scale.

W Chain, a leading innovator in blockchain-powered financial infrastructure, has officially launched W Swap, a next-generation decentralised exchange (DEX) built natively on the W Chain blockchain. This milestone is complemented by the debut of WAVE, the protocol’s high-yield incentive token designed to power sustainable decentralised finance (DeFi) growth. Since its launch on 17 June 2025, W Swap has experienced remarkable early momentum, with daily transaction volumes surging from just 100 pre-launch to 4,520 on 22 June, signalling strong community interest and early adoption across key markets.

Built on W Chain’s high-throughput, low-fee Layer-1 blockchain, W Swap employs an Automated Market Maker (AMM) model that enables efficient liquidity provision through initial pools such as WCO/USDT and WAVE/WCO. Liquidity providers earn 0.3% trading fees, distributed via LP tokens, while also benefiting from elevated WAVE emissions—particularly during the early adoption phase. The architecture has been optimised for ease of use, capital efficiency, and accessibility for both experienced DeFi users and newcomers.

The WAVE token, which launched staking functionality on 20 June 2025, is the cornerstone of W Chain’s incentive structure. Users can stake LP tokens to earn WAVE in addition to trading fees, or directly stake WAVE tokens themselves to earn projected yields of up to 700% APR. Within weeks of launch, WAVE is already trading at 300% above its initial price, a reflection of strong market demand and effective tokenomics. Furthermore, 10% of all emissions are directed to a protocol-owned treasury, ensuring “unruggable” liquidity and reinforcing W Chain’s commitment to long-term sustainability.

Community feedback has been overwhelmingly positive, highlighting both the usability and performance of the platform. One early user, Danny, commented: “W Swap is just what we’ve all been waiting for—fast, slick, and a great starting point for W Chain’s journey into DeFi. We can now bridge from Ethereum and BNB Chain into W Chain and farm using the Wave platform. Currently, the APRs are massive, and we’re building our bags. A great project with a great team to work with.”

Echoing this sentiment, Gary, another active community member, noted: “W Swap has a clean, crisp UI that is simple to understand—even the less experienced members of the crypto community will find it easy to use. We’re still early in terms of adoption, but as more trading pairs are added, the platform will become even more appealing. W Swap ties seamlessly into the farming, bridge, and pooling functionalities.”

Looking ahead, W Chain is preparing to roll out its next major protocol upgrade—WSWAP v4—targeted for Q4 2025 to Q1 2026. This version will incorporate significant enhancements inspired by Uniswap v3, including concentrated liquidity, multiple fee tiers, non-fungible LP positions, and a time-weighted average price (TWAP) oracle. These core upgrades will vastly improve capital efficiency and integration capabilities across the DeFi stack.

Beyond the foundational features, WSWAP v4 will introduce several advanced innovations aimed at extending functionality and improving user experience. These include a Hooks Framework to enable programmable logic during swap and liquidity events; a Dynamic Range Auto-Rebalancer offering passive and active liquidity strategies; and an ERC-20 LP Token Wrapper to allow NFT LP positions to be used across lending, staking, and yield farming applications. Additional improvements include fee booster mechanics, staking integration with WAVE and WCO, gas-optimised smart contracts, cross-pool routing, and DAO-governed extensibility for secure and flexible protocol evolution. Real-time analytics, profit tracking, and integration with W Chain’s native oracle layer will further enhance transparency and capital decision-making.

With a clear roadmap, rapidly growing user base, and high-performing DeFi primitives already live, W Chain is positioning itself as a leading platform for decentralised finance in Asia-Pacific and beyond. The combination of strong infrastructure, user-centric design, and high-yield opportunities provides a compelling case for liquidity providers, yield farmers, and developers alike.

For more information on 
W Swap – https://app.w-swap.com/#/
Wave – https://wave.w-chain.com/

About W Chain

W Chain is an innovative leader in blockchain technology, developing secure and efficient payment solutions for businesses and individuals worldwide. With a focus on combining the benefits of both public and private blockchain networks, W Chain Network is creating a next-generation platform for global payments, smart contracts, and decentralised finance.

For more information, please visit, https://w-chain.com/

MBG is backed by real-world assets and built for deep liquidity

Following the announcement of its utility token MBG, MultiBank Group has released additional details on the broader ecosystem in which the token will operate. MBG is designed as a transactional, incentive, retention and inter-business settlement layer for its four-pillar ecosystem combining traditional finance infrastructure, real-world asset tokenization, and digital asset trading.

MultiBank Group holds 17 reputable regulatory licenses across five jurisdictions with an unblemished track record. Today, the Group reports $35 billion in daily trading volume on average. The launch of MBG marks its entry into the Web3 space, supported by a structured ecosystem and real-world business operations. MBG is gaining traction and welcoming Waitlist sign-ups ahead of its July 2025 launch.

Overview of MBG’s ecosystem

MBG will be used across the Group’s four-pillar ecosystem for transaction fees, staking, rewards, and internal settlements. The MBG ecosystem consists of four key components, each leveraging the token for internal utility:

• MultiBank TradFi: The Group’s existing brokerage and trading platform, offering access to traditional financial markets, predominantly for FX and CFDs.
• MultiBank.io: A digital asset platform offering spot trading, derivatives, and future DeFi integrations.
• MultiBank.io RWA: A tokenization system for real estate assets, starting with a $3 billion agreement involving Dubai-based MAG Lifestyle Development, that offered its Ritz-Carlton Residences, Dubai, Creekside at Keturah Resort, and Keturah Reserve for the launch in 2025.
• MEX Exchange: An upcoming institutional-focused crypto Electronic Communication Network (ECN), designed for large-scale liquidity aggregation. Scheduled to launch in 2025.

MBG’s growing community

MBG is a thriving community supported by MultiBank Group’s 2 million+ customers across the ecosystem, with 35% identified as existing crypto enthusiasts. The community is actively engaged across MBG’s social channels:

Twitter: 267k https://x.com/multibank_io
Telegram announcements: 183k https://t.me/MultiBankio_Announcements
Telegram: 118k https://t.me/MultiBank_io

A project backed by deep liquidity and built on rrust

Unlike many tokens that launched, MBG is issued by a profitable financial group with no debts, regulated operations and audited financials. Key stats include:

• In 2024, the Group reported a net income of US$275.4 million, a balance sheet of US$607 million, and a valuation that stands at US$29 billion
• MultiBank Group holds 17 regulatory licences with an unblemished record
• MultiBank Group serves over 2 million customers across 100 countries, with 35% identified as crypto users

MEX Exchange, MultiBank Group’s institutional ECN bridging TradFi and crypto, will be the first institutional platform for inter-bank trading in emerging markets, providing deep liquidity and setting a precedent for crypto adoption among institutions. It is projected to handle over $460 billion in daily volume within five years and designed to provide prime brokerage and instant client settlement. Today MEX Exchange is valued at US$23 billion.

Bridging сrypto and the real world

The MBG token is built around a high-value real estate tokenization initiative, beginning with a $3 billion agreement in Dubai. The developments include The Ritz-Carlton Residences, Dubai, Creekside at Keturah Resort, and Keturah Reserve that will be offered to customers on day 1 of launch. A $10 billion pipeline to tokenize real estate is underway, positioning the ecosystem as one of the most ambitious RWA platforms to date.

Investors can enter these premium tokenised real estate assets with as little as $50 and earn up to 8% APY, paid out daily. Around the clock, tokenholders can access fractionalized real estate through a secured marketplace, merging tangible assets with digital liquidity.

The MBG token is central to how the platform functions – powering yield payouts, staking, discounted trading fees, and broader ecosystem access.

Deflationary buyback and burn programme

MultiBank Group has implemented a token burn model to manage the circulating supply that links circulation directly to platform activity. Rather than follow a fixed schedule, the burn rate adjusts in line with actual usage, grounding the model in adoption rather than speculation. The program aims to remove as much as 50% of the total token supply over time. The structure is designed to reinforce value and maintain a healthy token economy tied to real participation. Key features:

• Cumulative burn of up to $440 million over the next four years, and a goal to burn up to 50% of the total supply
• Burn schedule is based on ecosystem usage, with no fixed timeline for maximum reduction
• The burn program aims to remove as much as 50% of the total token supply over time

The road ahead

The Group has outlined a phased rollout for its crypto infrastructure:

• 2025: MultiBank.io Derivatives Exchange Launch – expanded to include Silver and Gold contracts with leverage; MultiBank.io Swap/Convert Product Launch; MBIO Mission Center Launch; Spot Trading APIs Launch; MultiBank Chain (Private Testnet); OTC Desk Launch; MBIO RWA Launch; MBG TGE Campaign, CoinGecko, CoinMarketCap, other major data aggregators, and DEXs Listing; Referral Program; Launch of the MBG Ecosystem Utility Token
• 2026: Derivatives Trading APIs Launch; OTC Retail Portal Launch; Spot FX and Metals Trading, Margin and Credit Trading, Social Trading; Learn to Earn Academy; ECN TradFi; Advanced AI
• 2027: OTC Institutional Portal Launch, Multibank.io Escrow/CLS Service Launch; MBIO Pay (Caas-PSP); Multibank.io Derivatives Social Trading; Multibank.io Pay Launch
• 2028: MultiBank Group Stablecoin (MUSD) Launch; Introduction of the first Web3 Trading Platform for institutional markets; Launch of the Multibank Smart Chain
• 2030: Projected ecosystem trading volume exceeding $540 billion per day

For more details, refer to the MultiBank Group White Paper and Pitch Deck.

Summary

With MBG, MultiBank Group is building a multi-layered financial ecosystem for both retail and institutional customers, grounded in regulated operations, asset-backed products, and digital infrastructure. The token is designed to serve a specific function across platforms rather than functioning as a speculative vehicle.

Join the MBG waitlist — backed by a financial giant bringing TradFi to crypto.

For more information on MBG or MultiBank Group, visit:
MultiBank TradFi: https://trade.multibankfx.com/
MEX Exchange: https://mexexchange.com/
MultiBank.io: https://multibank.io/
MultiBank.io RWA: https://multibank.io/rwa
MBG Token Promo Sites:
https://token.multibankgroup.com/en/early-access-mbg/waitlist
https://token.multibankgroup.com/en

Alltius today launched its Agentic AI Suite, a comprehensive platform for building, deploying, training, and maintaining sophisticated AI agents geared to redefine the financial services landscape. This will enable financial institutions to move beyond basic automation and into an era of intelligent, proactive, and personalized customer experiences.

“We are at an inflection point in the history of artificial intelligence and its application in the enterprise,” said Vibs Abhishek, CEO and Founder of Alltius. “With our Agentic AI Suite, we are providing the engine for the next generation of financial services—one that is not only more efficient and automated but also more human-centric and intelligent. We are giving our clients the power to build their future, today.”

The Alltius Agentic AI Suite: An Architecture of Innovation

The Alltius platform is a complete, end-to-end ecosystem for agentic AI built for insurance, banking, lending, wealth management and FinTechs, providing a powerful suite of tools that work in concert to deliver unparalleled performance:

• PULSE Analytics: Go beyond dashboards to unearth deep, actionable insights into customer behavior, operational efficiency, and new product opportunities. See the “why” behind the data and anticipate customer needs before they arise.
• INTERACT Human AI Interface: A seamless, intuitive interface for your employees, enabling them to collaborate with AI agents through voice and text. Deploy your AI agents for your customers or your employees with one click integrations across voice, email, chat, custom APIs and more. 
• ACT Multi-agents Library: Leverage a powerful library of over 50+ pre-trained AI agents, specifically designed for the complexities of the financial services industry. Pick and choose from AI agents for banking, financial services, wealth management, insurance agency management, call center management and more. 
• FLOW Engine: Design and deploy complex, intelligent workflows with an intuitive, natural language interface. The Flow Engine empowers both business and technical users to create powerful automations that drive efficiency and innovation.
• KNO Store: Seamlessly connect your existing systems and data sources with over 100+ integrations and almost all kinds of documents. Provide your AI agents with the context they need to deliver truly personalized experiences.

The Alltius Difference

Alltius is the only platform that works with the latest AI research, combining the strengths of multiple research methodologies, top-tier and proprietary LLMs, and a world-class team of researchers to deliver best-in-class performance for every financial services task. With over 20 years of experience developing these advanced systems, our team has an unparalleled understanding of the nuances and complexities of enterprise AI.

Transforming Financial Services: Use Cases for a New Era

The Alltius Agentic AI Suite is already transforming a wide range of financial services use cases, including:

• Loan Origination and Underwriting: Automate lending operations, data collection, risk assessment, and decision-making to accelerate the loan lifecycle and improve accuracy.
• Intelligent Customer Servicing: Provide 24/7, personalized support to your customers, resolving their inquiries instantly and freeing up your human agents to handle more complex issues.
• Automated Claims Processing: Streamline the claims process from first notice of loss to final settlement, reducing costs and improving claimant satisfaction.
• Proactive Wealth Management: Empower financial advisors with AI-powered insights and recommendations, enabling them to provide proactive, data-driven advice to their clients.
• Next-Generation Compliance and Fraud Detection: Leverage AI to monitor transactions in real-time, identify suspicious activity, and ensure compliance with all regulatory requirements.

A Partnership for Success: End-to-End Hand-Holding

Alltius is more than just a technology provider; we are a partner in your success. We offer comprehensive consulting services, providing end-to-end hand-holding to ensure that you get the most out of our platform. Our team of experts will work with you every step of the way, from initial strategy and design to implementation, training, and ongoing optimization.

The Alltius Agentic AI Suite is available now. To begin your journey to the future of financial services, visit alltius.ai.

About Alltius

Alltius is a pioneering agentic AI platform for financial services. By combining deep financial domain expertise with advanced natural language processing and system integration capabilities, Alltius enables financial institutions to deliver exceptional customer experiences while dramatically reducing operational costs. Headquartered in Irvine, Alltius serves insurance companies, banks, and wealth management firms globally.

Alltius’ innovative no-code platform allows businesses to seamlessly create, train and deploy AI assistants, even within a day. Alltius’ platform has military grade security with SOC2 Type 2 and ISO certifications, ability to handle enterprise scale volumes of more than 500K queries per day, no hallucinations with native AI technology & 1/10th operating costs with self-hosted solutions.

Alltius is a product created by experts who have taught AI and NLP at Carnegie Mellon and built large scale systems at places such as Amazon, Google and Meta and used by billions

Decentralized AI Framework Gains Recognition for Expanding Access to Compute Power.

The digital economy has witnessed transformative platforms that fundamentally changed resource sharing: Grab revolutionized transportation, Airbnb transformed hospitality, and Shein disrupted supply chains. Now, a Vietnamese technology company is redefining the next frontier-computational power sharing itself.

Hyra Network has been officially named “Technology Startup of the Year” at the prestigious 2025 Globee Awards for Technology, marking a watershed moment for decentralized artificial intelligence infrastructure. This recognition validates an ambitious vision that could reshape how the world builds, owns, and benefits from AI technology. (Official winner list available at: globeeawards.com/2025-winners-technology)

The Vision Behind Innovation

This breakthrough platform is researched and developed by Hyra Tek JCS (Vietnam) and operated by Hyra Tek Smart Solution L.L.C (UAE). Hyra Network’s mission is to democratize computational power by activating billions of idle devices and transforming everyday users into AI infrastructure providers.

Hyra Network serves as the flagship platform alongside Hyra AI, creating an unprecedented model where computational resources are shared across distributed networks rather than concentrated in centralized data centers. If ride-sharing optimizes vehicle utilization and home-sharing maximizes property efficiency, then computational sharing unlocks vast processing power lying dormant in smartphones, computers, and IoT devices globally.

Global Recognition for Excellence

The Globee Awards represent the technology sector’s highest honor, with winners selected by over 100 seasoned professionals, including C-suite executives, venture capitalists, and industry analysts.

“This honor transcends our company – it validates the entire movement toward democratized AI infrastructure,” said Mr. Jonh Tran, Founder of Hyra Network. “We’re witnessing global acknowledgment that the future of AI belongs not to centralized monopolies, but to communities that collectively own and benefit from these powerful technologies.”

Pioneering Community-Powered AI

At its core, Hyra Tek’s innovation centers on Hyra AI, one of the world’s first Train-to-Earn platforms. This system allows users to convert personal devices into active AI training nodes, earning rewards while contributing to advanced model development. The Layer-3 blockchain architecture supports high-throughput, low-latency workloads, enabling scalable AI training and inference at the network’s edge.

The economic model creates a virtuous cycle: participants provide computational resources, earn tangible rewards, and simultaneously advance AI capabilities that benefit the broader ecosystem.

Global Impact

Today, the Hyra ecosystem spans more than 205 countries, powering a global network of approximately 2.5 million connected devices, including 700,000 active online nodes and over 1 million KYC-verified users. With strong community engagement across Southeast Asia, Latin America, and Africa, Hyra delivers more than 360,000 teraflops of distributed computing power and supports a growing base of enterprise clients, now serving over 10 paying customers. This real-world adoption reaffirms Hyra’s core belief: that distributed, permissionless infrastructure can drive meaningful innovation while remaining truly open and accessible to all.

As artificial intelligence and DePIN technologies gain momentum, Hyra positions itself at the technological vanguard, architecting the foundation for a more equitable, intelligent, and inclusive AI future.

About Hyra Network

Hyra Network is a decentralized artificial intelligence infrastructure platform developed by Hyra Tek JCS (Vietnam) and operated by Hyra Tek Smart Solution L.L.C (UAE). The platform enables individuals and enterprises to contribute unused computational resources from personal devices to support AI training and inference at scale. Leveraging a Layer-3 blockchain architecture, Hyra Network supports high-throughput, low-latency workloads across a global network of connected devices. Its ecosystem includes Hyra AI, one of the first Train-to-Earn platforms, and currently spans over 205 countries with 2.5 million connected devices. Hyra Network aims to increase accessibility to AI infrastructure through a community-powered, distributed model.