vai

On October 19, the team from the fully homomorphic encryption company Zama announced that users participating in the incentive distribution platform Merkl on the Zama confidential vault can now directly receive cZAMA rewards in the Zama app. The rewards are settled through confidential transfers, and both the positions and the reward amounts remain confidential from deposit to distribution.In September 2026, Zama launched confidential access to 16 selected vaults on the lending protocol Morpho and initiated the first batch of confidential incentives with Merkl, where rewards are calculated based on encrypted balances. When depositing into eligible vaults, users can activate additional incentive options, granting viewing permissions for Merkl to read encrypted balances to calculate rewards without disclosing personal positions. This permission is enabled by default on the deposit interface, and deposits not shared with Merkl do not qualify for incentives.Confidential vaults accumulate based on the user's confidential vault shares and can be claimed during the ongoing activity; existing public vaults with confidential access have rewards collectively obtained by the vault packaging contracts, which are distributed after the activity ends. Zama recommends claiming confidential activities at the end to reduce multiple on-chain fees. Taking the Flowdesk Confidential High Yield USDT vault on Morpho as an example, the related activity ended on October 12, and eligible depositors can now claim their rewards. Users need to check their earnings in Merkl using their wallet and complete the claim in the Zama app by connecting the same wallet; the amount received will not be publicly disclosed on-chain.

Robinhood announced that its social trading network is now open to all eligible users in the United States. CEO Vlad Tenev stated on October 13 that this feature is built into the app's information feed, centered around users' trades and performance. Users can follow other traders, view shared account activities, and initiate eligible trades directly from posts through the Quick trade feature. Tenev mentioned that Robinhood Social is now available to everyone, allowing users to share charts, profits and losses, and portfolio allocations, as well as receive news and company updates.This feature was previously tested starting on March 18, initially for 1,000 customers who attended the HOOD Summit, followed by an additional 10,000 in the subsequent weeks. The early version provided verified trading posts, daily and yearly profit and loss statistics, and the ability to place orders manually from the information feed; news feeds are part of future planning. Sharing is selective: trades will not be automatically published, and customers can choose which eligible trades to publish, as well as disable trading statistics in their profiles.Access and publishable content are restricted. Social is only available in the United States, and Robinhood excludes licensed securities professionals, individuals associated with regulated financial institutions, and the company's own employees. Its help page indicates that futures and multi-leg options trades currently do not meet publishing criteria. Tenev stated that Social will be coming to Robinhood's desktop trading platform Legend in the coming weeks.

Google has started to roll out Gemini 4 Argon to Google AI subscription users, applicable on platforms including Google Antigravity, the Gemini app, and the AI mode in Google Search. Gemini 4 Argon is launched at a promotional price of $2 per million input tokens and $10 per million output tokens.

B.AI platform announces that Anthropic's latest Haiku model, Claude Haiku 5.5, is officially launched, with API and Web Chat both open for experience. This model is designed for high-frequency calls and latency-sensitive workloads, achieving a better balance between speed, cost, and intelligence, suitable for well-defined sub-agent tasks. The model defaults to adaptive thinking, supports five levels of reasoning intensity, has image understanding and tool invocation capabilities, and supports 1M Token ultra-long context with a maximum output of 128K. It is applicable in scenarios such as classification, routing, information extraction, summarization, real-time customer service, code-assisted retrieval, and interface automation. From now on, developers and users can directly call Claude Haiku 5.5 through the B.AI platform.

On October 7, WEEX Exchange announced the launch of the "Wall Street Tycoon × Monopoly" event. Users can sign up to receive dice, complete tasks such as trading real-time U.S. stocks, contract trading, consuming stock tokens, and inviting friends to earn more dice; rolling the dice to move on the board will unlock random rewards such as BTC / ETH / USDT / WXT / stock tokens / bonuses / deductions / position airdrops / points / dice, and users can also participate in U.S. stock prediction contests daily to win more points and share a prize pool of one million USDT. Event duration: October 7, 00:00 to November 3, 23:59 (UTC+8).Among them, from October 7 to 11 is the TOKEN2049 themed phase. Users who sign up to participate in the event on-site at TOKEN2049 Singapore can receive a limited VIP experience card. In addition, participants can also visit the WEEX Alpha Suite to join roundtable forums, AMAs, AI trading experiences, and Mini Trading House interactions, receiving exquisite merchandise and custom souvenirs.

On September 30, the B.AI platform announced the official launch of OpenAI GPT-6.1 Sol, with API and Web Chat both open for simultaneous experience. As an upgraded version of GPT-6 Sol, this model is aimed at complex software engineering, computer operations, and professional workflows, officially positioned to be close to GPT-6 Astra in intelligent agent programming, computer usage, and professional work, with the standard input/output token price being about 1/5 of Astra.The model supports 1.05M token ultra-long context and a maximum output of 128K, providing 5 adjustable reasoning strengths, and supports prompt caching and Responses API multi-agent delegation (Beta). From now on, developers and users can directly call GPT-6.1 Sol through the B.AI platform. You are welcome to log in to chat.b.ai/chat to experience it immediately.

On September 29, B.AI announced that Anthropic's latest Sonnet model, Claude Sonnet 5.5, has officially launched on the platform. As a faster and lower-cost supplement to Claude Opus 5.5, Claude Sonnet 5.5 boasts an output speed increase of over 30%, with a maximum reduction in single-task costs of about 30%, aimed at clearly defined software engineering, everyday agent workflows, and professional document creation scenarios.This model supports image and PDF comprehension, can generate reports, presentations, spreadsheets, and UIs, and offers five adjustable reasoning intensities with a 1 million token context window, with no additional charges for long contexts. Currently, both the B.AI API and Web Chat have been synchronized and opened, allowing developers and users to directly access this new generation Sonnet model that balances speed, cost, and long context capabilities.

Binance Wallet announced that Binance Alpha has launched Doppler Finance (XDP). Users holding at least 230 Binance Alpha points can claim 1666 XDP airdrops on the Alpha event page, on a first-come, first-served basis, until the airdrop pool is exhausted or the event ends.If the event has not ended, the point threshold will automatically decrease by 5 points every 5 minutes. Claiming the airdrop will consume 15 Binance Alpha points, and users must confirm their claim on the Alpha event page within 24 hours, or it will be considered forfeited.

B.AI platform announces the official launch of two efficient models in the OpenAI GPT-6 series, GPT-6 Sol and GPT-6 Luna, with API and Web Chat available simultaneously.The entire series supports 1.05M ultra-long context and 128K output, natively supports image understanding and Computer Use, and provides 6 adjustable inference strengths. Among them, GPT-6 Sol is aimed at complex software engineering and Agent workflows, scoring 68.8% on DeepSWE v1.1, with single-task costs reduced by about 80%; GPT-6 Luna focuses on high-frequency large-batch loads, scoring 66.6% on DeepSWE v1.1, with input as low as $0.10/1M Tokens.Starting today, developers and users can directly call the two models through the B.AI platform.

On September 23, the B.AI platform announced that the first model of Anthropic's newly released Claude 5.5 series, Claude Opus 5.5, is officially online, with API and Web Chat both open for simultaneous experience. This model is designed for long-cycle Agent programming and complex knowledge work, achieving Claude Fable 5.1 levels in most tasks, while reducing operating costs by about 40% compared to Opus 5.The model supports 1M Token ultra-long context and up to 128K output, demonstrating leading capabilities in Computer Use and multi-step automation scenarios. From today, developers and users can directly access Claude Opus 5.5 through the B.AI platform. You are welcome to log in to chat.b.ai/chat to experience it immediately.

The B.AI platform announces that the latest preview model of Moonshot AI, Kimi K2.8 Preview, is officially launched, providing developers with simultaneous access to API and Web Chat dual-end experiences.This model is specifically designed for programming and Agent workflows, supporting 1M (1,048,576 Tokens) ultra-long context and multimodal inputs of text, images, and videos. Its overall performance closely approaches the flagship Kimi K3, with further improvements in thinking efficiency and coding capabilities compared to K2.7 Code, and it supports three adjustable reasoning intensity levels: Low/High/Max.Starting today, developers can directly invoke Kimi K2.8 Preview through the B.AI platform.

DeepSeek stated in the V4.1 Flash open-source announcement that if there are 2000 GPUs and storage clusters, they can contact DeepSeek directly.V4.1 Flash is still open-sourced under the MIT License, without additional licensing conditions based on company revenue or business scale. Companies with resources can deploy it themselves, and contacting DeepSeek is more like a direct technical collaboration for ultra-large-scale scenarios. DeepSeek has not yet announced a formal private deployment product or paid service.In the past, after DeepSeek open-sourced the model, deployment adaptations were mostly completed by inference frameworks and hardware vendors such as SGLang and TensorRT-LLM. This time, the official team has proactively provided direct contact information for large-scale deployments and stated that they will continue to support the open-source community in adapting inference for new models.

Quantus founder and CEO Yuvi Lightman posted on X that the anti-quantum blockchain Quantus mainnet will go live on September 9 (Wednesday). The only way to acquire tokens at the launch is through mining, and liquidity and tradability will be available a few weeks later.The team has merged Immunefi security work into the public code repository and completed all fixes from the audit competition. The chain supports ML-DSA-65 and ML-DSA-87, binding each signature to its context according to FIPS 204 standards, failing and closing multi-signature transactions that cannot be decoded, and strengthening the network stack to resist denial-of-service attacks. The team has completed a full-process drill on the global rehearsal mainnet covering the Americas, Asia, Europe, and Australia, producing signed binaries through an automated pipeline, deploying multi-region bootnodes and dedicated RPC nodes as infrastructure as code, and reducing indexer memory usage from over 90% to about 26%.In addition, Quantus has launched a registration page, completed a website redesign, Keystone hardware wallet supports multiple accounts, and released CLI v2.1.3 and miner v4.0.2. Quantum & Privacy Day will be co-hosted with NEAR in Singapore on October 8, coinciding with Token2049.

The Compound Foundation announced the launch of the Institutional Market, a lending market exclusively for whitelisted institutions, supporting borrowing of ETH, wstETH, WBTC, and cbBTC with USDC, with a maximum loan-to-value ratio of 87%. On its first day of launch, the market was oversubscribed, with participation from institutions such as DeFi Saver, K3/Nexo, KPK, and Yearn. The entry threshold is a deposit of 100,000 USDC, and Compound has reserved 200,000 USDC as supplier incentives, which will be proportionally distributed to whitelisted participants within three months, with a supply cap of 20 million USD.This market is the first product to be launched from the 52 million USD plan approved by the Compound DAO in August, described by the foundation as the Institutional Comet built on v3.5, independent of the DAO-funded V4 roadmap. Compound stated that v3 has been running for four years without any incidents. Currently, Compound's total locked value is 1.53 billion USD, with 1.42 billion USD on the Ethereum chain, ranking sixth among lending protocols on DefiLlama, with a 23% growth over the past 30 days.The Compound Foundation was relaunched on August 17 with a budget of 52 million USD, with a team from Coinbase, Anchorage, NEAR, and Maple.

The payment system X Money tweeted that its services are now available to all Premium and Premium+ subscribers with a U.S. account.

B.AI platform announced yesterday that its free model lineup is fully ready, covering six top models including GLM-5.3-Flash, DeepSeek-V4-Flash, DeepSeek-V4-Flash-Vision-Exp, Hy3 (Tencent Hunyuan), MiMo-V2.5 (Xiaomi), and Qwen3.8-Flash, all available to users with zero barriers for free.Among them, the debuting GLM-5.3-Flash is the "Ox Alpha" that has attracted widespread attention. As the first native multimodal model in the GLM-5 series, it features a total of 320B parameters and 18B active parameters, using a sparse + linear attention hybrid architecture and supporting 1M ultra-long context, combining rapid response with strong reasoning capabilities, making it the cost-effective choice for high-frequency calling scenarios. This free lineup covers various scenarios including pure text, multimodal, code generation, and Agent scheduling. Whether developers or deep AI users, they can access cutting-edge large model capabilities at zero cost through the B.AI platform, simply by logging in at chat.b.ai/chat to start an efficient workflow.

SpaceXAI announced the expansion of the availability of the Grok robot, adding support for SuperGrok Plus, Cursor Pro+, and Cursor Teams packages.

According to The Defiant, Kieran Warwick, co-founder of the Web3 game project Illuvium, stated that the studio has significantly reduced costs over the past six months, and the team recently agreed to further salary cuts. According to him, the adjusted funds can support operations for more than 12 months. He did not disclose the extent of the salary cuts, the number of remaining employees, or the cash reserves.Warwick mentioned that the existing team is almost entirely focused on the MMO game Overworld, with most members having worked together for over three years. In addition to the MMO, there are plans to launch the Celebrations series of Illuvium Beyond, Beyond Chaos, Primal Tactica developed by independent studios, and Loadout, which will allocate tokens for staking contracts, within the year. Primal Tactica will participate in Steam Next Fest in October, and Illuvium will receive most of the revenue.Data from DefiLlama shows that Illuvium's protocol fees and revenue over the past 30 days are approximately $9,781, with a treasury of about $2.78 million. ILV is priced at approximately $3.35, with a market capitalization of about $27.74 million.

On August 20, the B.AI platform DeepSeek-V4-Flash zero-threshold free access event continues. B.AI upholds the vision of "Inclusive Computing Power" and is committed to building a globally leading one-stop AGI infrastructure, continuously lowering the threshold for AI use, allowing developers and Web3 users to easily experience, build, and deploy Agentic applications.Relying on the one-stop AGI infrastructure, B.AI aggregates cutting-edge model matrices from around the world, with built-in intelligent routing to optimize costs and efficiency; stably supporting high-throughput scenarios such as millions of long contexts and Agent workflows; elastic API routing balancing production-level high availability and extreme cost-effectiveness; Web2/Web3 dual channels bridging crypto assets and fiat currency settlement. Currently, DeepSeek-V4-Flash is continuously available for free on both Web and API ends, with $0 access to top model capabilities.

AI programming tool Cursor announces the launch of the Origin code hosting feature, which is now available for early testing to all paid users starting today. This feature supports the creation of new code repositories, bidirectional synchronization with GitHub repositories, and includes pull requests, code browsing, and AI Agent integration capabilities. Users can complete code reading, questioning, modification, and PR merging all within the same interface.The synchronized GitHub repositories support bidirectional comment synchronization. Comments in Cursor will sync to GitHub, and replies on GitHub will also be sent back to Cursor in real-time. Origin also integrates applications such as Vercel, Depot, and Buildkite, allowing PRs to automatically generate preview deployments. Cursor indicates that more native Agent features will be launched in the future. Enterprise users can opt out of this testing program through their administrators.