Search “Agents”

At the Android Dev Summit, Google launched the Bring Your Own Agent feature for Android Studio Rabbit 2. Through the open Agent Client Protocol, developers can run more than 40 coding agents simultaneously within the IDE, including Claude Agent, Codex, and Google Antigravity, with direct access to project structure, the emulator, and Logcat. Google says that in internal tests, after agents rendered Compose previews in the terminal, they completed UI tasks 77% faster and cut token usage to as lit

Matched fieldsSummary、Body

English text is being prepared.

Matched fieldsSummary、Body

Xinyan Robotics has released Bubbo 1, an embodied home agent built on its self-developed Bubbo OS Family Agent. It will go on sale starting October 29 through official channels including JD.com, priced at 9,499 yuan, with a discounted price of 7,999 yuan for the first 1,000 orders. Bubbo is equipped with 13 categories and 26 sensors, 6 degrees of freedom, and two expression screens. It features built-in camera-based heart rate detection and enables data sharing with the Cece smart ring. Its conv

Matched fieldsSummary、Body

Madhulika Srikumar and Claire Leibowicz of Partnership on AI write that current agent frameworks cannot reliably generate the key observability records needed for safe deployment. Together with colleagues from academia, civil society, and industry, the authors studied the four most widely used agent frameworks and found that signals such as persistent Agent IDs and human approval records are missing, making it difficult to determine which agent did what and whether it exceeded its authority. The

Matched fieldsSummary、Body

r-35B-A3B-Preview, which uses a MoE architecture with 35 billion total parameters and about 3 billion activated per token, and supports a thinking mode. On the 49 tasks of the web Agent benchmark WindTunnel, its median execution time was 1.52 seconds, lower than the 3.21 seconds of the Jev + Mercury 2.5 combination, with success rates of 98.6% and 95.9% respectively. On code benchmarks, it scored 73.6 on SWE-bench Verified and 51.7 on Terminal-Bench 2.1, and the team said it will subsequently la

Matched fieldsSummary、Body

earning on top of the preview version, with improved coding and defensive cybersecurity capabilities. It scores 19.7% on FrontierCode 2.0 and 48% on the Artificial Analysis Coding Agent Index, roughly on par with GLM-5.3, but uses an average of 47% fewer tokens per task; on Terminal-Bench 4.0 it improves by 13 percentage points to 41%. The model supports three reasoning modes, uses a modified MIT-based Mistral Ecosystem Open-Weight license, and is priced via API at $0.70 per million input tokens

Matched fieldsSummary、Body

ning turn-taking in conversational AI: the model must switch among three actions—replying, holding the floor, and listening to silence—relying on adaptive timing, VAD, and semantic analysis to determine whether the user has finished speaking, rather than depending on a fixed silence timer. When Eleven v4 Turbo is paired with ElevenAgents, the median latency from request to audible speech is 142 ms. Engineering teams can measure conversational fluency using response latency and interruption rate.

Matched fieldsSummary、Body

. Customers no longer need to purchase hardware devices upfront, but must commit to at least 36 months. The service is based on its latest firewall platform R82.20 and includes an agentic console for managing firewall devices and services, Force hardware firewalls, SandBlast's AI threat protection, and instant security updates. Check Point says the move is intended to address emerging threats brought by AI agents and to lower the barrier for enterprises to upgrade with a predictable monthly fee.

Matched fieldsSummary、Body

neration world model that can generate interactive environments in real time and collaborates with Gemini 3.8 Flash through a division of labor: the latter observes the visuals, decomposes tasks, and adjusts plans. Leiphone.com analyzes its technical challenges from three aspects—dynamic language condition injection, Agent causal hallucination, and dual-model asynchronous collaboration—and notes that DeepMind has not yet disclosed the network architecture or scheduling implementation of Genie 4.

Matched fieldsSummary、Body

R&D tasks had reached the AL4 automation level, meaning AI can autonomously complete most steps based on high-level goals. The prospectus also disclosed four previously unreported Agent safety incidents, one of which involved a model in a cybersecurity test that, because the environment was not disconnected from the network as required, used saved passwords to obtain third-party system administrator privileges and access external data, a process that lasted 10 to 34 hours. During the same period

Matched fieldsSummary、Body

nforcement learning (RL), and post-training an as-yet-unreleased editing model based on Qwen 3.8 27B. In early work, training Llama 3.3 70B on more than 250,000 samples was enough to compete with top models on editing tasks at the time. The Figma agent is now available to all users; in testing, typical tasks are about 20% faster, in expert comparisons it was better than or on par with the previous production agent in 62% of cases, user ratings were 33% higher, and undo operations were 60% fewer.

Matched fieldsSummary、Body

activated parameters, trained from scratch on 3,800 NVIDIA Grace Blackwell GPUs in its own data center in Europe. The company says it scores 48% on the Artificial Analysis Coding Agent Index, roughly on par with GLM-5.3, but consumes 47% fewer Tokens per task on average. The model is released under a modified MIT license: personal, academic, fine-tuning, and conventional commercial use are free, while cloud providers offering MaaS resale need to obtain a separate commercial license; online API

Matched fieldsSummary、Body

Baseten has become Postman's recommended inference partner for enterprise agents, responsible for providing production inference and post-training support for agents built on Postman Agent Stack. Baseten says its seven model APIs released in 2026 all ranked in the top two in Artificial Analysis's throughput and latency rankings, with Beam 1.0, DeepSeek V4.1 Flash and others taking first place. Harvey used Baseten to post-train a 27B open-source model, matching closed-source frontier models on le

Matched fieldsSummary、Body

ncent announced an upgrade to the WorkBuddy Enterprise Edition open strategy, and used Buddy One to connect Tencent Docs, Tencent Meeting, Tencent Drive, and ima, sharing the same Agent foundation along with a unified account and credit pool. Dowson Tong disclosed that since its launch in March this year, WorkBuddy Enterprise Edition has been deployed in more than 50 industries, covering 50% of the Fortune 500 China companies and 60% of China's top 100 enterprises, with more than 50,000 open pla

Matched fieldsSummary、Body

sixth. Carbon mainly improves complex coding tasks and cybersecurity capabilities, scoring 88.2% on SEC-bench Pro, higher than Argon's 83.4% and OpenAI GPT-6 Astra's 85.4%; however, it still trails Claude Opus 5.5 on the Coding Agent Index, Terminal-Bench, and Frontiercode. Currently, Carbon is available only to selected users such as cybersecurity defense experts, and will be opened to developers, enterprises, and consumers only after passing the U.S. government's voluntary pre-review process.

Matched fieldsSummary、Body

hts are already available on Hugging Face and via its API. Mistral says the model's performance is comparable to that of the open-source model GLM-5.3, but it consumes 47% fewer tokens per task, and it ranks first among open-source models on Artificial Analysis's Coding Agent Index. Pierre Stock, Mistral's Vice President of Science, said that open weights mean organizations can deploy locally in their own environments to identify, coordinate, and respond to major security incidents in real time.

Matched fieldsSummary、Body

Epsilon's Agentic Platform team used Amazon Nova Act together with Amazon Bedrock AgentCore Browser to turn intent-driven UI testing into a multi-tenant platform service layered on top of its existing Playwright regression suite. Test steps are written in natural language, executed by Nova Act over a CDP WebSocket, with per-step screenshots and pass-rate statistics, and results stored in Amazon S3. The solution provides managed temporary browser sessions, AWS IAM-based authentication, and cross-

Matched fieldsSummary、Body

er and model adaptation, with pilot programs in Hangzhou, Ningbo, Wenzhou, and Jiaxing. Alibaba Group, Zhejiang University Science Park Development Co., Ltd., and China Telecom Zhejiang Company simultaneously released lists of empowerment items, among which Alibaba provides more than 70 mainstream models with 1 million free tokens each, and China Telecom Zhejiang Company provides the highest-tier membership of the TeleAgent Xingchen super intelligent agent free of charge until December 31, 2026.

Matched fieldsSummary、Body

tries. At the event, it released the Buddy One cross-app AI collaboration system, which connects WorkBuddy, Tencent Docs, Tencent Meeting, Tencent Drive, and ima to share a common Agent foundation, and simultaneously launched Tencent Drive to serve as unified storage. The enterprise edition also upgraded its security and observability capabilities and released the WorkBuddy Managed Agent (WMA), allowing enterprises to codify workflows into Skills and have them run 24/7 in the cloud; the ecosyste

Matched fieldsSummary、Body

weekly review. The new process is divided into five steps: pulling transaction records, reconciliation, data extraction, report generation, and sending. More than 30 USB security keys are operated by RPA through a USB key dock, and the Agent uses Python to unify the transaction record formats of various banks and automatically reconcile balances and transaction counts. The project took about 9 days from on-site demo to launch, and ran in parallel with manual work for about 2 weeks after launch.

Matched fieldsSummary、Body