Gate News message, April 21 — OpenClaw, an open-source AI Agent platform, released v2026.4.20 on April 20, introducing significant updates to default model switching, plugin contract handling, and gateway storage optimization.
The release switches the default model for Moonshot’s built-in channels, web search, and multimodal understanding to Kimi K2.6, while maintaining backward compatibility with Kimi K2.5. Moonshot now allows thinking.keep = “all” exclusively on Kimi K2.6; other Moonshot models or requests with fixed tool_choice parameters will have the field automatically removed. The update adds tiered pricing for cost estimation with built-in pricing for Kimi K2.6 and K2.5, enabling direct token usage billing.
A critical fix reverts the strict ID contract introduced on April 14, which required plugin info.id to match registered slot IDs and caused third-party context engine plugins like lossless-claw to be rejected on every round. BlueBubbles text delivery default timeout increased from 10 to 30 seconds, and macOS 26 Tahoe now prioritizes Private API even when disabled, preventing silent message loss.
On the operations side, Cron now separates runtime execution state into jobs-state.json while keeping jobs.json for git-friendly task definitions. Sessions storage enables default entry limits and time-based trimming on startup, preventing memory buildup in gateway and executor. The Onboarding wizard redesigned security instructions with yellow warning banners and sectioned lists, added loading animations for initial model directory loading, and included placeholders in API key input fields. Security was further tightened for Gateway WebSocket broadcasts, device pairing permissions, and workspace .env injection blocking for OPENCLAW_* keys.
Disclaimer: The information on this page may come from third parties and does not represent the views or opinions of Gate. The content displayed on this page is for reference only and does not constitute any financial, investment, or legal advice. Gate does not guarantee the accuracy or completeness of the information and shall not be liable for any losses arising from the use of this information. Virtual asset investments carry high risks and are subject to significant price volatility. You may lose all of your invested principal. Please fully understand the relevant risks and make prudent decisions based on your own financial situation and risk tolerance. For details, please refer to
Disclaimer.
Related Articles
Meta rolls out AI employee keyboard-tracking training during layoffs amid a wave of job cuts
Meta announced on April 21 that it will install a new tracking tool called “Model Capability Initiative” (MCI) on employees’ computers in the United States, to capture employees’ mouse movements, clicks, keyboard input, and periodic screenshots of their work screens, for training AI agent models capable of autonomously carrying out job tasks. This move came against the backdrop of Meta cutting roughly 2,000 employees this year, sparking strong reactions from employees.
MarketWhisper7m ago
Volkswagen to Deploy AI Agents in Chinese Market From H2 2026
Volkswagen to deploy China-exclusive AI in cars from H2 2026, with Horizon Robotics, enabling navigation, reservations, and parking; aims to boost software leadership and price competitiveness in China, launching 20+ new electrified models in China and 50 globally by 2030.
Abstract: Volkswagen announced at a Beijing event on April 21 that it will embed an AI assistant into its Chinese-market vehicles starting in the second half of 2026, using a China-exclusive electronic architecture developed with Horizon Robotics. The AI system surpasses conventional voice assistants by performing multi-step tasks such as locating and reserving restaurants, providing navigation, and coordinating parking. Volkswagen China CEO Ralf Brandstätter described the platform as part of the automaker's strategy to close the software gap with local rivals and to mitigate pricing pressures in the world's largest EV market. The plan accompanies an aggressive product roll-out, including more than 20 new electrified models in China and about 50 globally by 2030, signaling a renewed emphasis on software-enabled differentiation in its global electrification push.
GateNews10m ago
NeoCognition Emerges from Stealth with $40M Seed Round Led by Cambium Capital
Gate News message, April 22 — NeoCognition, an AI startup founded by Ohio State professor Yu Su, has emerged from stealth with a $40 million seed round led by Cambium Capital and Walden Catalyst Ventures. Vista Equity Partners also participated in the round, with angel investors including Intel
GateNews50m ago
Moonshot AI Releases Kimi K2.6 Open-Source Model with Advanced Agent Workflows
Gate News message, April 22 — Moonshot AI released its open-source Kimi K2.6 model on April 20, introducing new capabilities for coding, front-end generation, and multi-agent orchestration. The model is designed to support complex enterprise workflows through advanced Agent Programming and Deep
GateNews2h ago
Visa Launches Intelligent Commerce Connect for AI Agent Payments
Visa unveils Intelligent Commerce Connect, enabling AI-initiated purchases via a single integration on the Visa Acceptance Platform, with payment initiation, tokenization, authentication, and controls; pilots with partners, rollout planned this year.
Abstract: Visa introduces Intelligent Commerce Connect, a service that lets AI agents initiate purchases through a single integration on the Visa Acceptance Platform. It supports payment initiation, tokenization, authentication, and spending controls for both Visa and non-Visa cards, helping businesses accommodate AI-driven shopping. The service is currently piloted with partners such as Aldar, AWS, Diddo, Highnote, Mesh, Payabli, and Sumvin, with broader rollout planned later this year.
GateNews3h ago
Google launches Deep Research Max: supports MCP and can access enterprise private data
According to a Google DeepMind official blog announcement, Google launched a new generation of autonomous research agents, Deep Research and Deep Research Max, on April 21, 2026. They are built on Gemini 3.1 Pro, as the official release following the preview version provided through the Interactions API in December 2025. Both agents are now available in the paid tiers of the Gemini API in the form of a public preview, and new startups and enterprise users on Google Cloud will be able to connect gradually.
The two variants are positioned differently: interactive vs asynchronous depth
Google categorizes the two agents by use case: Deep Research
ChainNewsAbmedia9h ago
