搜索“Agent”

eepSWE v1.1, a test of long-horizon software engineering, against 74.2% for Anthropic PBC’s Claude Opus 5.5. OpenAI Group PBC’s GPT-6 Astra sat a 10th of a point further back. The agentic coding interface from Google Antigravity now runs directly inside the Gemini app for macOS, which means a chat can turn into a coding session without a jump to a separate tool. Google did not mention a Windows version. The app was one part of a wider release today. Koray Kavukcuoglu, chief AI architect at Googl

匹配字段正文

ished a blog post calling for “speed limits” on technological development. OpenAI’s Sam Altman supported it in a rare sign of solidarity between the rival companies, following its agents autonomously hacking Hugging Face in July—as well as Australian government websites and several other entities, with activity continuing into September. What’s next for Muse Meta hasn’t had a Hugging Face moment—yet. For now, its Muse assistants are actively helping people complete tasks on their computers quick

匹配字段正文

igger trend we’ve seen across the AI industry. As criticism of AI and the companies building it mounts — from the anger over data center construction to the questions raised by AI agents making unwanted incursions — Big Tech leaders have pivoted to speak about AI’s ability to cure cancer, for example. Gemini 4 Argon competes with frontier models like OpenAI’s GPT-6 Astra and Meta’s new Muse Wonder. Google had the edge among AI companies for most of 2025, but advancements from other labs left it

匹配字段正文

s” in coding performance, Google is letting users “switch between chatting and coding” within the Gemini app for macOS. Users will specifically have direct access to Antigravity’s agentic coding interface in the consumer client for building apps, web experiences, and more. Google Antigravity + more Top comment by Mckillio The irony of using AI to save on processing power. But this is where I'm probably most excited, the backend improvements/fixes that will benefit end users. Glad to see this fin

匹配字段正文

Argon turned months of internal engineering into weeks Google shared some internal numbers with the launch, though they're self-reported. Software engineers using Argon increased agent-driven code submissions by more than 35 percent, while code rollbacks dropped by nearly a third, the company says. In the Chromium project, Argon helped the Chrome team complete an architectural refactor across hundreds of files within the 40-million-line codebase. Work that would have taken months was finished i

匹配字段正文

models; Speech recognition Data science; Data centers; Data analytics Data management; Synthetic data The company’s new releases point to an AI market becoming more personal, with agents and models completing tasks, but enterprises still aren’t ready for this new phase of AI. October 13, 2026 Meta introduced models on Monday that target personal and professional agentic tasks, capitalizing on the momentum of its successful Muse personal agent, released last month. Muse Wonder and Muse Wonder Ful

匹配字段正文

ld models; Real2Sim → Reconstructing physically calibrated simulation twins from robot camera feeds, egocentric human video, or hand-drawn paper sketches; Physical understanding → Agentic multimodal hindsight labeling including detailed understanding of dexterity, dense value functions and progress estimation; Research and self-improvement → Action improvement via simulation and iterative self-improvement on hardware Unlocking robot control with Argon 1. Spot mobile manipulation: Putting in the

匹配字段正文

our hardest engineering problems. Over the past few weeks, software engineers have been shipping high-quality code faster than ever. Argon in Antigravity has helped teams increase agentic code submissions by more than 35% while reducing code rollbacks by nearly a third. Untangling browser feature code in Chrome: Argon helped the Chrome team separate core webpage handling from a 6 million line legacy folder of browser features held within Chromium’s 40 million-line codebase. Piping multiple clean

匹配字段正文

ld models; Real2Sim → Reconstructing physically calibrated simulation twins from robot camera feeds, egocentric human video, or hand-drawn paper sketches; Physical understanding → Agentic multimodal hindsight labeling including detailed understanding of dexterity, dense value functions and progress estimation; Research and self-improvement → Action improvement via simulation and iterative self-improvement on hardware Unlocking robot control with Argon 1. Spot mobile manipulation: Putting in the

匹配字段正文

our hardest engineering problems. Over the past few weeks, software engineers have been shipping high-quality code faster than ever. Argon in Antigravity has helped teams increase agentic code submissions by more than 35% while reducing code rollbacks by nearly a third. Untangling browser feature code in Chrome: Argon helped the Chrome team separate core webpage handling from a 6 million line legacy folder of browser features held within Chromium’s 40 million-line codebase. Piping multiple clean

匹配字段正文

and shipping great products. Check out the Gemini API Developer Guide for model details and setup instructions; Read the Google Antigravity blog for details on using Argon in our agent-first development platform; Read the Android Studio blog to see how Android Developers can start building; See how you can use Gemini 4 Argon for robotics Developer tools; Gemini models Related stories Developing Gemini 4 Argon responsibly and safely Gemini 4 Argon: Built to solve your hardest problems EmbeddingG

匹配字段正文

progress more often and more effectively. Argon also brings major improvements in coding, so we’re making it easier to switch between chatting and coding with Google Antigravity's agentic coding interface now available directly in the Gemini macOS app. It’s also easier than ever to use Gemini to get things done right from the main chat window, with a simple toggle from chat to tasks. When you choose tasks, Gemini can work for you even when you’re not at your computer, helping you get more done a

匹配字段正文

On this page Today, we release LFM2.5-24B-A2B, a 24B-parameter reasoning model with just 2B active parameters, built for fast, private agents on-device. It is the largest and final model in the LFM2.5 family, which now spans two orders of magnitude: from LFM2.5-230M to LFM2.5-24B-A2B. LFM2.5-24B-A2B is competitive with Qwen3.6-35B-A3B, a model with 50% more total and active parameters, on five of eight agentic and tool-use benchmarks, scoring 48.09 on LiveClaw versus 38.19 and 27.69 on Automatio

匹配字段正文

BREAKING 🔥: Meta is rolling out Muse Wonder and Muse Wonder Full on the Meta Model API platform! Muse Wonder - State-of-the-art at real-world coding, built for long-running tasks. Plans the work, calls tools in parallel, hands parts off to sub-agents, and recovers from errors so complex jobs finish.

匹配字段正文

Muse Wonder is Meta's most capable model yet, designed for complex personal and professional agentic tasks. It delivers frontier coding performance, advanced computer use capabilities, and handles long-horizon work to tackle developers' most demanding tasks. https://t.co/zOJ2jWuvSt

匹配字段正文

Its ahead of GPT-6 Astra and Claude Fable 5.1 on long-running coding, legal work and business automation. The widest gap comes on SWE-Marathon, an ultra-long engineering test, where Full scores 57.5 against 46.3 for Opus 5.5. Full also leads DeepSWE 1.1, a long-horizon coding test, 79.0 to 74.2, in what Meta says were independent third-party runs. On Harvey’s legal agent benchmark, which credits only tasks finished in full without hallucination, Full scores 13.1, more than triple Opus 5.5’s 4.2.

匹配字段正文

Meta is so back. in each of these benchmarks Muse Wonder beats both Opus 5.5 and GPT-6 Astra The widest gap comes on SWE-Marathon. this one tests whether an AI agent can build an entire software project on its own, such as a C compiler or a Slack clone, over 2 to 10 hours, and counts a task only if every hidden check passes. And muse wonder full scores 57.5 against 46.3 for Opus 5.5.

匹配字段正文

e for users on Meta’s first-party API, while Muse Wonder Full is currently only available for Muse Code users. Meta notes that each model is designed for personal and professional agent use Congratulations to @AIatMeta, @finkd, and @alexandr wang on the release! Key Takeaways: ➤ Both models exhibit a material jump in overall Intelligence Index score. Muse Wonder Full (max) improves on Muse Spark 1.3 (max) across all ten Intelligence Index evaluations, gaining 8 points to 56. The largest gains we

匹配字段正文

@tabContentCard buttonType = "default" [[tabs]] label = "WorkBuddy" name = "TabInfoCard" [[tabs]] label = "元宝" name = "TabInfoCard" [tabs.data] logo="yuanbao" label = "元宝" role = "Agent算法负责人" nickname = "Jim" content = "接入 Hy3 后，用户使用元宝 Agent 感受到的是：Excel 不用反复调格式，数据分析结果直接可用，网页工具接入即跑。覆盖信息查询、数据处理、文档办公、生活决策、网页制作五大场景的内部评测，给出数字支撑： 文档生成综合分提升 +7%，网页制作与自动化脚本提升 +6%。 " [[tabs]] label = "ima" name = "TabInfoCard" [[tabs]] label = "Marvis" name = "TabInfoCard" [[tabs]] label = "QQ浏览器" name = "TabInfoCard" [[t

匹配字段正文

—— 短于 0.18s 姿态只展开一半读成瞬移，\n 长于 0.35s 玩家会拿它赶路而不是救命。\n 写不出上下界的数，说明这个数还没想清楚。\n4. 写自动化验收脚本（puppeteer 无头截图 + 断言），至少覆盖：\n 实例数/三角面对账、草与配色、面板控件计数、描边外壳闭合性、\n 角色步态、特效可见度、镜头位移上限（单步 来完成这个游戏\n Agent 任务提示词：卡牌构筑肉鸽游戏（类杀戮尖塔）从 0 到可发行\n\n 一、角色与总目标\n\n你是一名独立游戏全栈开发 Agent，需要独立完成一款 卡牌构筑 Roguelike（Deckbuilding Roguelike） 游戏的设计、开发、美术、测试与打包。参考标杆是《杀戮尖塔》（Slay the Spire），但玩法数值和内容必须原创，不得抄袭原作的卡牌名称、角色形象与美术资产。\n\n 最终交付物：一个可以双击运行、无需安装 Godot 或任何运行时的 Windows 桌面版游戏（.exe），内容量达到\"完整单章体验\"，且已通过自动化测试验证可玩性与平衡性。 \n\n你必须自主推进到底，不要在中途反复询问细节

匹配字段正文