model

B.AI announced that the Kimi K3 model is officially launched on its API platform, seamlessly integrated after the release by @Kimi_Moonshot. As the world's first open-source 3T level (28 trillion parameters) model, Kimi K3 has made significant breakthroughs in long-range programming, complex reasoning, and knowledge work, and natively supports visual capabilities and 1M ultra-long context. As a model aggregation service platform, B.AI has prioritized opening its API interface, allowing developers to directly experience the capabilities of the new generation model through B.AI. B.AI will continue to provide efficient and stable interface services to promote the rapid implementation of cutting-edge AI technology and empower developers' innovation.

According to Xinhua News Agency, as the 2026 World Artificial Intelligence Conference and the High-Level Meeting on Global Governance of Artificial Intelligence are about to be held, the new generation model Kimi K3 was released on the 16th, with a parameter scale of 28 trillion. This is currently the largest open-source model in the world by parameters, marking a new step in the development of artificial intelligence models in our country.Beijing Moon's Dark Side Technology Co., Ltd., which released the model, introduced that Kimi K3 natively supports visual understanding, has a context window of 1 million tokens, and is optimized for complex task scenarios such as software engineering, knowledge work, deep research, and multimodal understanding, further enhancing the processing capability of large models for complex tasks. It is understood that in evaluations, Kimi K3's comprehensive intelligence level is close to that of the world's leading closed-source models. A relevant person in charge of Moon's Dark Side stated that in the process of training the new generation foundational model Kimi K3, they used a self-developed underlying model architecture and accumulated a complete set of scientific model training methods.

According to the Financial Times, informed sources reveal that the Chinese AI unicorn company Moonshot AI plans to release a new large language model, Kimi K3, in the near future. This model has between 20 trillion to 30 trillion parameters, making it the largest AI model in China by parameter scale, and its performance is expected to surpass the flagship model Claude Opus 4.8 from Anthropic in mainstream benchmark tests (industry speculation suggests its parameter count is around 15 trillion to 20 trillion).Unlike the currently mainstream closed-source and expensive cutting-edge large models in the United States, Kimi K3 will be available as an open-weight model for users to download and modify for free, which may create competitive pressure for leading American labs like OpenAI and Anthropic. Currently, due to the rising service fees for large models in the U.S. (for example, Anthropic has announced a 50% price increase for Opus 4.8 in September), some overseas companies have begun to shift towards using more cost-effective Chinese open-source models.In terms of the capital market, informed sources indicate that Moonshot AI is preparing for a new round of financing, with the latest valuation expected to reach approximately $31.5 billion. Meanwhile, the valuations of other AI giants in China and the U.S. are also rising; DeepSeek is starting a new round of financing with an estimated valuation of about $71 billion, while Anthropic and OpenAI have reached valuations of $965 billion and $852 billion, respectively, in their latest round of financing. In response to the aforementioned release and financing rumors, Moonshot AI has currently declined to comment.

Digital asset management company CoinShares stated that Open USD directly challenges Circle by generating revenue from the distribution of stablecoin reserves to partners, undermining the distribution economic model of USDC.Open USD consists of over 140 companies, including BlackRock, Coinbase, Mastercard, Stripe, and Visa. The stablecoin is expected to launch in the second half of 2026. CoinShares noted that despite the competitive pressure posed by Open USD, the existing liquidity and integration foundation of USDC may be difficult for new entrants to replicate.

According to a report by Reuters, the Cyberspace Administration of China announced on Wednesday (July 15) that Apple's edge-generative AI service "Apple Intelligence" has successfully completed filing in China and is approved for use on iPhone devices in the Chinese market. The official announcement has not yet specified a concrete launch date.According to informed sources, Apple's AI service will integrate the AI model capabilities of Baidu and Alibaba. Alibaba subsequently confirmed in a statement that its "Tongyi Qianwen" (Qwen) large model will be integrated into the Apple Intelligence experience for users in China on iOS, iPadOS, macOS, and visionOS. Apple has not yet responded to this.

a16z published a blog post stating that as traditional financial institutions accelerate their exploration of blockchain technology, the market generally believes that the future will see a comprehensive integration of DeFi (Decentralized Finance) and TradFi (Traditional Finance), forming a new financial model through the combination of decentralized finance and institutional distribution systems.However, the reality may not be so. The core motivation for traditional financial institutions to adopt blockchain is not to embrace decentralization, but to value its commercial benefits in reducing costs, improving settlement efficiency, expanding distribution channels, and optimizing customer relationship management.What is more likely to emerge in the future is a new type of "programmable financial infrastructure" based on underlying blockchain technology, optimized for institutional needs, rather than a simple integration of traditional finance and DeFi. Institutions are selectively absorbing certain technological capabilities from DeFi and modifying them according to their own regulatory, risk management, and operational requirements.For example, atomic settlement can reduce counterparty risk, shared ledgers can lower back-office reconciliation costs, programmable funds can automatically execute processes such as interest payments, margin management, and corporate actions, and automated market-making models are also being applied to on-chain foreign exchange and tokenized asset pricing.At the same time, the native DeFi features of open access, anonymity, and trustless execution often conflict with institutional requirements for compliance, control, and accountability. Therefore, cases such as JPMorgan's institutional blockchain project, BlackRock's and Franklin Templeton's tokenized funds, are essentially not traditional finance entering DeFi, but rather using blockchain technology to improve existing financial business processes.In the future, the blockchain industry will have two development paths: on one hand, enterprises and financial institutions will continue to promote the implementation of blockchain infrastructure that meets regulatory requirements, expanding the industry scale through applications such as stablecoins, tokenized assets, and on-chain settlements; on the other hand, open networks will continue to play the role of a source of innovation, continuously generating new financial primitives and market mechanisms, providing technical reserves for future institutional infrastructure.TradFi and DeFi are not in competition but are developing together in different directions. Traditional finance may not fully adopt the DeFi model but will gradually adopt parts that suit its own needs. The true integration may ultimately occur at the underlying blockchain network level, rather than one side replacing the other.For developers, the key is not to chase all markets simultaneously but to clarify the target audience: for institutions, products need to be built around compliance, risk control, and long-term business processes; for open networks, there is a need to continue exploring innovation, liquidity, and network effects. The future financial system may operate on blockchain infrastructure, but the most important innovations may still come first from open networks.

B.AI announced that the entire series of GPT-5.6 models has officially landed on Web Chat. Previously, the B.AI API had already integrated the three models: GPT-5.6 Sol, GPT-5.6 Terra, and GPT-5.6 Luna. Starting today, users no longer need to configure the API; they can log in to directly experience the complete capabilities, including chat Q&A, content creation, code writing, and complex reasoning. This full opening marks the all-scenario implementation of the new generation of models within the B.AI ecosystem, providing "out-of-the-box" intelligent services for developers and ordinary users. In the future, B.AI will continue to lower the barriers to using cutting-edge AI technology, facilitating efficient creation and decision-making.

According to CoinDesk, a wholesale digital market report supported by the UK Treasury lists Ripple as one of the core participants, planning to transition repurchase agreements (repo), UK government bonds (gilts), and funds from the regulatory sandbox to the real market within the next 12 months, with an expected annual increase of about £33 billion in output and £14 billion in tax revenue for the UK economy over the next decade.The report proposes a hybrid architecture that overlays a permissioned institutional network on top of a public chain, using BlackRock's tokenized money market fund BUIDL on Ethereum as an example, while also noting that the finality risk of settlements brought about by public chain restructuring remains to be resolved. The report also mentions Ripple's acquisition of Hidden Road (renamed Ripple Prime) and Santander UK's use of the Ripple blockchain for cross-border payments, suggesting that traditional finance and crypto institutions are converging.

According to CoinDesk, as the competition for tokenization in the capital markets heats up, the Securities Transfer Association (STA) recently submitted a letter of opinion to the U.S. Securities and Exchange Commission (SEC), warning that stock tokens issued by third-party organizations may undermine market integrity and calling on regulators to prioritize support for tokenized securities authorized by publicly listed companies in future rule-making.The STA represents several Wall Street transfer agents, whose members believe that true tokenized stocks should be formally authorized by the issuing company and recorded in the official shareholder register, rather than created as "packaged" token products by independent platforms.The association pointed out that third-party stock tokens may confuse investors about the actual rights they hold and expose them to risks related to platform credit, custody, and operations, without establishing a direct legal relationship with the publicly listed company. Therefore, any innovative exemptions, pilot projects, or permanent regulatory frameworks for tokenized securities should prioritize the issuer-supported model.The STA also urged the SEC to reform the existing Direct Registration System (DRS), arguing that the current U.S. securities custody system is inadequate to meet the real-time transfer and settlement needs of on-chain securities, and suggested that regulators collaborate with the Depository Trust & Clearing Corporation (DTCC) to optimize the digital securities infrastructure.Currently, the global market for tokenized stocks, valued at approximately $2 billion, is primarily dominated by third-party models, including products launched by Ondo Finance and Kraken, while organizations like Securitize and Figure adopt the issuer authorization model.

According to CoinDesk, Jurien Timmer, the Global Macro Director at Fidelity, stated that Bitcoin is approaching the bottom support line of its "power law model" constructed based on the complete price history of Bitcoin. This long-term support level is currently around $58,000, while the current price of Bitcoin is approximately $62,700, gradually nearing this area, or may have already entered the "accumulation zone" defined by the model.However, Jurien Timmer did not confirm that the market has bottomed out, but believes that Bitcoin may not experience a rapid rebound immediately. Instead, it is expected to fluctuate around the long-term support line for several months. Short-term capital has previously exited, and the flow of funds has rotated from Bitcoin to gold, and then from gold to the semiconductor sector. Currently, the market's pursuit of hot sectors is still mainly concentrated in the semiconductor field.

B.AI announced that it has officially integrated OpenAI's latest generation GPT-5.6 full model series into its API network. Starting today, developers can directly call three newly positioned models through B.AI's single interface, including GPT-5.6 Sol, which focuses on complex reasoning tasks, GPT-5.6 Terra, which is suitable for everyday workloads, and GPT-5.6 Luna, which emphasizes rapid response and high cost-effectiveness. This integration aims to provide developers with more flexible and stable computing power support in scenarios such as building intelligent agents, optimizing workflows, and exploring AGI applications.

According to The Information, internal memos show that SpaceXAI and Cursor plan to release their jointly developed first AI model as early as today (July 8). It is reported that to further enhance operational efficiency, the two companies postponed the release of the model earlier this week. The report indicates that the new model has the capability to process information quickly and aims to compete with advanced large models such as Anthropic's Opus 4.8 and OpenAI's GPT 5.5.Data shows that Elon Musk's SpaceX acquired xAI (now renamed SpaceXAI) in February of this year, and subsequently acquired the well-known AI programming assistant Cursor's parent startup Anysphere in June for $60 billion in an all-stock transaction to strengthen its position in the high-profit enterprise-level AI code generation sector. Additionally, SpaceX has performed strongly since its initial public offering on June 12, benefiting from Nasdaq's revised rules for newly listed companies, and was officially included in the Nasdaq-100 index this Tuesday.

The Berachain mainnet PoL Next has officially launched, and the contract upgrade marks the first step in transitioning from the original BGT model to a more simplified and sustainable incentive system centered around sWBERA. The complete transition is divided into two phases: on July 7, the contract changes will take effect, and WBERA will begin to be issued as a replacement for BGT; on July 8, the hard fork will completely stop BGT production, and the old mechanism will officially close.After the transition is complete, BGT LST Reward Vaults and BGT LST incentives will gradually be phased out in the coming days. The official will provide migration tools for all users on the Portfolio page of the Hub, but the migration will not be executed automatically; users will need to manually exchange BGT and BGT LSTs for sWBERA. Early APR may fluctuate, with an estimated maximum increase of up to 3 times the original. The hard fork time is July 8 at 16:00 UTC (12:00 PM EDT).

Vitalik Buterin posted that there is a winner in the AI anonymous recognition challenge experiment he previously initiated. In this experiment, he wrote the Chinese version of EIP-7503 (Zero-Knowledge Wormhole) in 2024, manually corrected after using Qwen 2.5 for local translation, in an attempt to hide his identity.The article states that although the prose style was effectively blurred, the AI successfully identified it by detecting unique thinking habits in mathematics and algorithm explanations (such as specific numerical examples and logical chains), completely bypassing its obfuscation strategy that only targeted prose.

According to the Financial Times, the UK's Financial Conduct Authority (FCA) has warned that regulation of artificial intelligence in the financial sector is facing an "arms race" and urged for new powers to regulate ChatGPT, Claude, and Gemini.

According to a report from Hongxing News, an investigation report leaked from a scam park in Myanmar shows that global telecom fraud is accelerating towards an "AI industrialization + cross-border encrypted payment" system. The scam network completes fund circulation through cryptocurrency and uses automated tools based on large models for multi-language script generation, identity disguise, and emotional manipulation.According to the investigation analysis, these systems extensively utilize OpenAI's ChatGPT and Google's Gemini at the functional level to support "scaled social fraud." The funding side achieves rapid laundering and transfer through on-chain payments and cross-border channels, forming a dual structure of "AI customer acquisition + encrypted settlement," which gives the fraud industry a high degree of automation and transnational diffusion capability. In addition, Starlink, owned by Musk, has become the number one network service provider for the scam park in Myanmar, with American ISP providers carrying nearly one-fifth of the park's traffic.In response to the related accusations, OpenAI stated that the behavior of scammers using ChatGPT is highly similar to that of ordinary users, making identification difficult. However, they have banned about 100,000 suspicious accounts monthly through behavioral pattern recognition and risk control systems. Google stated that its AI models have safety barriers in place and emphasized its commitment to "responsible AI development" to limit the tools' use for fraud and other illegal purposes.