AIsaAIsaStudio English 打开 AIsa Studio

Google 开放 Antigravity /teamwork 多智能体功能,展示 Gemini 4 在数学与系统工程的成果

原标题:Advancing science & engineering with Antigravity’s /teamwork

Google 发布 Antigravity 的 /teamwork 多智能体功能及 Gemini 4(Argon、Carbon、Hydrogen)系列成果,包括完整证明 2013 年信息论的 Courtade–Kumar 猜想并用 Lean 4 形式化为 5000 万行代码、自主生成约 180 万行 Lean 形式化 Weil 猜想,以及 9x9围棋引擎不到八天提升约 2600 Elo 达到 1885 Elo 等系统成果。该功能现已在 Antigravity 中开放,配合 /boost 和 Deep Review 使用,并向 1 万名学术研究者提供六个月最高档免费额度。

为什么值得看

文章汇总了 /teamwork 在证明、化学、系统等领域的具体成果与开放方式,读者可对照评估多智能体工作流在自身研究中的用法。

阅读原文 ↗11月14日 AI 日报

本页的摘要、推荐理由与解读出自 AIsa Studio 的 AI 风向智能体,原文版权归原作者所有。

深度解读

发生了什么

Google Research 宣布将 Antigravity 的 /teamwork 多智能体功能向研究者和开发者开放,并同步发布一批由 Gemini 4 系列模型(Argon、Carbon、Hydrogen)与多智能体工作流协作完成的成果。过去几周里,Google 内部团队与学术合作机构借助这些工具,在数学、科学与复杂系统工程的公开难题上取得了进展,覆盖信息论、形式化验证、分子结构发现、内核优化,以及 Google 生产基础设施的多项改进。

本次开放的工具包括 Antigravity 中的 /teamwork(与 /boost 及其他智能体并列)和 Deep Review(作为 /teamwork 的文档审阅模式实现)。Google 还开放了申请,向 1 万名学术研究者提供六个月最高用量档位的免费使用,并表示形式化证明模式等更多功能将在未来几周进入 Antigravity。

关键细节

/teamwork 的设计针对当前语言模型的两个已说明的局限。其一,单次生成的结果不可预测,因此系统将并行探索与专门的证伪器配对——如 Lean 4 证明检查器、未经修改的软件测试套件、执行模拟器、网页搜索和硬件计数器——自动验证候选方案。其二,模型在多日任务中会随上下文增长而丢失细节,因此 /teamwork「按构造划分上下文」:每个智能体只拿到所需文件和工具,协调者把任务交接给全新的后继者,而批评者和审计者从不含原始推理过程的干净上下文中验证结果。这些工作流以声明式、可自我改进的自然语言模式表达,并按提示动态路由。

帖子中列举的主要研究成果包括:

  • 由 Google、港中文、USC 和 CMU 研究者完成的 2013 年信息论 Courtade–Kumar 猜想的完整证明,已在 Lean 4 中形式化为 5000 万行代码。Gemini 首先找到对港中文团队广义猜想的反例,随后港中文团队将猜想归结为一个 14 变量优化问题,Gemini 4 Argon 划分参数空间并完成了大部分新的分析界和区间算术证书,还在无人介入的情况下把证明从均衡函数扩展到非均衡函数。
  • 在自适应数据分析中,运行于 Argon 上的多智能体框架独立证明并 Lean 验证了近乎最优的样本界,关闭了 2014 年以来的缺口。其他成果包括:首个在常数错误率下需要严格多于 n/2 次量子查询的 n 位布尔函数(解决 2014 年问题)、解决 STOC 2024 主要公开问题的确定性组合布尔矩阵乘法算法、稀疏网络距离近似的下界、打破稀疏性障碍的堆叠稀疏 sketch(2025 年问题),以及首个单遍流式割稀疏化算法(解决 FOCS 2022 遗留问题)。
  • 更多结果覆盖在线逆线性优化(最优悔憾、多项式运行时)、图灵机模型下低于 n log n 的整数乘法及一个与 GMP 在十亿位内逐位一致的 O(n log n) 网格算法、Tarjan 1985 年 Deque 猜想的进展,以及近似留一交叉验证在非光滑稀疏惩罚下的准确性保证。
  • 与 Google Cloud AI 合作的多智能体框架 LEAP 完全自主地完成了 Weil 猜想的端到端机器检查形式化——约 180 万行 Lean,跨 28800 次子智能体调用——并以 91.9 万余行 Lean 形式化了一般整数环上希尔伯特第十问题的一个近期证明,复现了带光滑外力的三维 Navier–Stokes 有限时间爆破的端到端形式化,还首次形式化了 1927 年以来未解的 Birkhoff 猜想和开放 50 余年的三维 Bochner–Riesz 猜想,后者在过程中发现并修复了近期一个非正式证明中的缺口。
  • 在核磁共振数据的分子结构判定上,30 个专用智能体组成带物理约束谱模拟器的自我纠错循环,在复杂基准分子上达到 95% top-1 与 100% top-3 准确率,几乎是单智能体表现的两倍。因果推断基准由 29 个智能体在 Argon 上运行 11.5 小时构建,把 13 张公开表格转化为 50 个任务,独立的审阅与审计智能体拦截了投机取巧、标记措辞提示并否决了幻觉脚本。

系统工程方面:基于 SNAP 微内核的低延迟内核旁路 PGAS 系统 FarArray,其无锁连通分量可扩展到一万亿边图(相对 TCP 键值存储吞吐提升 60–90 倍、加速 40–180 倍且 RAM 降低 6 倍);InstantEmbedding 将峰值 DRAM 降低最多 14 倍。为图神经网络定制的 TPU 内核实现最高 185 倍 HBM 减少。Argon 上的自主爬山团队从零训练出一个 9x9 围棋引擎,不到八天获得约 2600 Elo,达到 1885 Elo,自对弈加速 2300 倍。为期两周的 Argon 战役将 Zircon 微内核剩余 C++ 代码重写为约 80 万行安全 Rust 并在 arm64 硬件上启动;libgav1 被重写为约 17 万行 Rust——初期比原 C++ 慢约 4 倍,但在获得性能分析工具后,比原 C++ 和手写汇编在 x86_64 和 ARM 上快 20–30%。DuoDiscover 模式构建了 23 秒的代理评估器(用 Kendall 秩相关对照 7–32 分钟的全量硬件运行持续验证),在 17 个 JAXBench 内核上取得 4.09 倍几何平均加速,在 7 个 vLLM TPU-Inference Pallas 内核上取得 2.99 倍。生产应用包括超出最初 1 PiB 估算的全舰队内存优化,以及 Borg 智能体调度变更在部分硬件类型上减少 24% 的 TPU 任务抢占。

在 STOC 2027 之前,578 位作者(超过全部投稿的三分之一)选择使用 Deep Review;在持续进行的调查(N = 70)中,90% 的受访者认为反馈大部分或几乎全部属实,超过 30% 表示它发现了实质性证明错误。

为什么重要

这篇文章描绘了从「AI 作为助手」到「长时程智能体战役」的转变:智能体自行编写求解器、探索大搜索空间,并用机器可检查的标准验证每一步。数学成果中 Lean 4 形式化的一致运用尤为关键:一旦人类数学家确认形式化定理陈述与原始问题一致,「只有每个逻辑步骤都成立,证明才能编译通过」,这直接回应了形式化验证给同行评审带来的负担。

系统工程结果也并非只有基准分数:通过未经修改测试套件的内核旁路网络、能在真实硬件上启动的内存安全重写,以及 Google 内部的生产变更(Borg 调度、全舰队内存节省)。而 /teamwork 能将提示路由到从历史运行轨迹中自我改进的模式——「一个 harness 优化另一个 harness 的 token 效率」——表明多智能体框架本身,而不仅是模型,正在成为优化的对象。

对谁有影响

  • 数学、理论计算机科学和形式化验证领域的研究者,尤其是使用 Lean 4 或处理计算机辅助证明的人;STOC 2027 作者已可使用 Deep Review,普通用户每周可免费调用两次。
  • 可申请六个月最高用量档免费使用的 1 万名学术研究者;对外可用性是分阶段推进的,部分功能(如重写模式和形式化证明模式)尚未上线 Antigravity。
  • 从事分布式系统、内核优化和大规模代码迁移的系统与 ML 基础设施工程师,其中的模式(按依赖顺序重写并对照未修改测试套件、为昂贵硬件评估构建代理评估器)可直接借鉴。
  • 化学与实验室科学从业者(核磁共振结构判定工作流),以及需要构建基准的人(经审计的因果推断基准生成流程)。
全文翻译

以下翻译只覆盖原文的前一部分,其余内容请看下方原文。

在过去几周里,Google 团队与学术合作伙伴机构共同攻克了数学、科学和复杂系统工程领域中多年来悬而未决的开放性问题。借助 Gemini 4 Argon、Gemini 4 Carbon、Gemini 4 Hydrogen 以及多智能体工作流(如基于各种 Gemini 模型构建的 Antigravity’s /teamwork),这些合作在信息论、形式化验证、分子结构发现和内核优化方面取得了新的创新成果,同时也为 Google 带来了多项生产基础设施层面的提升。我们正将这些智能体工具在 Antigravity 的 /teamwork 中提供( alongside /boost 和其他智能体),以便研究人员和开发者能够将其应用于自己的工作中。

前沿研究和系统工程需要浏览海量文献、构建定制化的计算管道、针对物理测量结果进行测试,并验证每一个步骤。进步来自于一个交互式发现循环:研究人员提出问题并定义理论或物理约束,而长周期智能体则编写自定义求解器、探索庞大的搜索空间、测试候选设计,并检查长长的逻辑链。

扩展长周期推理能力

当前语言模型的两大局限性推动了 /teamwork 的设计。首先,模型生成具有不可预测性,导致单次尝试与找到正确答案之间存在差距。该系统通过将并行探索与专门的证伪器(例如 Lean 4 证明检查器、未经修改的软件测试套件、执行模拟器、网络搜索或硬件计数器)配对,自动验证哪些候选方案有效,从而弥合了这一差距。

其次,AI 模型在多日任务中往往会丢失细节,因为注意力分散在不断增长的上下文窗口中。为了防止这种情况,/teamwork 通过构建方式对上下文进行分区:每个专用智能体仅接收其在特定步骤所需的精确文件和工具。协调员将任务移交给新的后继者,而批评者和审计员则在干净的上下文中挑战和验证结果,而不包含原始生产者的推理过程。

Teamwork 将这些工作流表达为声明式的自我改进自然语言模式,并将每个用户提示动态路由到最适合该任务的模式。这些模式范围从分布式编码和迭代问题解决(作为 /boost 独立打包,这是一个从先前轨迹中优化的顺序求解和验证循环)到专门的研究工作流,如长证明和文档审查,所有这些都在 Antigravity 中可用。这使得团队能够从过去的运行记录中转录,跨不同框架模式设计自定义规范,并动态改进智能体框架。虽然许多这些功能在 Antigravity 中已经可用,我们将在未来几周提供更多更新,将所有功能提供给我们的外部 Antigravity 用户。继我们之前关于 Antigravity /teamwork 的文章之后,我们提供了由 /teamwork 新变体和 Gemini 4 模型赋能的开放挑战的新成果。

数学与理论计算机科学

在数学和理论计算机科学领域,研究人员与 Gemini 4 和 /teamwork 框架合作,共同开展大规模证明,并直接从开放问题陈述中生成完整的算法和证明。我们展示几个例子。

证明与验证 Courtade–Kumar 猜想

来自 Google、CUHK、USC 和 CMU 的研究人员利用 /teamwork 的 Long Proof 模式(在 Antigravity 中实现 Stellar Colosseum)来证明信息论中的 2013 年 Courtade–Kumar 猜想,涵盖了完全一般的情况。该结果证实,当随机比特串通过噪声信道时,没有任何将输入压缩为单个比特的方法能比“独裁函数”(即仅复制单个输入比特)保留更多信息。Gemini 首先发现了 CUHK 团队早期论文中广义猜想的反例。随后,CUHK 团队提出了一个新的候选函数,将猜想简化为一个 14 变量的优化问题,并提供了一个关键的解析下界。Gemini 4 Argon 和早期的 Gemini 模型随后对参数空间进行了分区,并开发了大部分新的解析边界和区间算术证书。他们还将证明从平衡函数扩展到非平衡函数,无需人工干预。团队在 Lean 4 中对整个证明进行了形式化,生成了 5000 万行代码。

解决优化、统计、量子计算和算法领域的开放问题

研究人员还直接向 Gemini 4 以及 Stellar Colosseum 和 /teamwork 等多智能体框架提出开放问题,收到了端到端的算法和证明(通常带有 Lean 4 形式化),作者对其进行了验证、简化和编辑以用于发表。在自适应数据分析中,运行在 Gemini 4 Argon 上的多智能体框架独立证明并经过 Lean 验证了针对从先前答案中自适应选择的统计查询的近最优样本界限,填补了自 2014 年以来一直存在的空白。这些界限产生了第一个近乎空间最优的算法,用于同时包含插入和删除的数据流,并能抵御自适应对手保持准确性。在量子计算领域,一个新的构造解决了一个 2014 年的开放问题,找到了第一个需要严格多于 n/2 次量子查询(在恒定误差下)的 n 位布尔函数,同时确定了跨对称函数的确切查询开销。

几项结果打破了处理和压缩大型矩阵及图形的长期障碍。直到 STOC 2024 的五十年间,布尔矩阵乘法和三角形检测的组合算法(避免快速矩阵乘法中的代数抵消)仅在立方时间上节省了对数因子。一个新的确定性算法解决了 STOC 2024 的主要开放问题,实现了其针对立方时间的目标加速。在高维几何中,新的下界解决了关于稀疏网络的开放问题,证明了即使使用额外的 Steiner 点和任意多的跳数,要在任何常数因子内近似 n 个点之间的所有成对距离,可能需要比 n 条边多得多的多项式级边数。在随机线性代数中,堆叠稀疏草图通过打破经典稀疏性障碍来解决一个 2025 年的开放问题,在不坍缩 r 维子空间任何方向的情况下将矩阵压缩至线性维度,每行仅需 √log r 个非零元素(比保留所有长度所需的 log r 少平方倍),并且几乎匹配新的下界。对于在动态边权重插入和删除下的加权图,Stellar Colosseum 解决了 FOCS 2022 留下的开放问题,开发了首个用于 切割稀疏化 的单遍流式算法,同时实现了近乎最优的内存、更新时间和重建时间。

我们的结果还涵盖优化、算术、数据结构和统计学。在优化领域,解决关于在线逆线性优化(随时间推断专家的隐藏目标)的开放问题,该方法提供了一种确定性算法,在实现最优遗憾值的同时,将运行时间从维度的指数级降低到多项式时间。对于 n 位整数乘法,多智能体活动简化了低于 n log n 的理论图灵机界限,而另一项努力给出了在一个二维网格上的新 O(n log n) 算法,其 C 语言实现在高达十亿位的范围内与 GNU 高精度算术库 (GMP) 逐位匹配。在数据结构方面,另一项结果在 Tarjan 的 1985 年 双端队列猜想(自调整 splay 树是否能在常数摊销时间内处理两端的插入和删除)上取得了重大进展,证明了操作开销的新上界,同时击败了所有有限迭代的逆 Ackermann 函数,并在 Lean 4 中进行了验证。最后,在高维统计学中,该方法证明了近似留一法交叉验证(避免从头重新拟合模型)对于像 弹性网 这样的非平滑稀疏惩罚仍然准确。

扩展形式化验证与手稿审查

手动检查计算机辅助的手稿给同行评审带来了沉重负担,细微的缺陷可能隐藏在漫长的计算中。在 Lean 4 中对数学进行形式化是将证明转化为代码,由计算机内核逐行检查。一旦人类数学家确认形式化的定理陈述与原始问题一致,如果每一步逻辑都成立,证明才能编译通过。

数学中许多最深刻的开放问题涉及物理波的传播、动力系统的演化以及多项式方程是否有整数解。直接基于非正式证明,我们与 Google Cloud AI 同事合作开发的 multi-agent 框架 LEAP,并与 Argon 配合,在这些核心领域取得了进展。

值得注意的是,/teamwork-LEAP 完全自主地生成了 Weil 猜想 的端到端机器检查形式化,这是关于有限域上多项式方程解计数的里程碑式成果,最终导致了 Pierre Deligne 1974 年的证明。该形式化跨越约 180 万行 Lean 代码,由 28,800 次子代理调用产生。该系统还合成了超过 919,000 行 Lean 代码,以生成关于一般整数环上 Hilbert 第十问题 最新证明的首次形式化。它还独立重现了端到端的形式化

查看原文

Over the past several weeks, teams across Google and academic partner institutions have tackled open problems in math, science, and complex systems engineering that had remained open for years. Using Gemini 4 Argon, Gemini 4 Carbon, Gemini 4 Hydrogen and multi-agent workflows such as Antigravity’s /teamwork (building on various Gemini models), these collaborations have produced new innovations across information theory, formal verifications, molecular structure discoveries, and kernel optimization, along with several production infrastructure gains at Google. We are making these agentic tools available in Antigravity’s /teamwork (alongside /boost and other agents) so that researchers and developers can apply them to their own work.

Frontier research and systems engineering require navigating vast literature, building custom computational pipelines, testing against physical measurements, and verifying every step. Progress comes from an interactive discovery loop where researchers frame questions and define the theoretical or physical constraints, while long-horizon agents write custom solvers, explore large search spaces, test candidate designs, and check long chains of logic.

Scaling long-horizon reasoning

Two limits of current language models motivate /teamwork's design. First, model generation can be unpredictable, leaving a gap between a single attempt and finding the correct answer. The system bridges this gap by pairing parallel exploration with dedicated falsifiers (e.g., the Lean 4 proof checker, unmodified software test suites, execution simulators, web search, or hardware counters) to automatically verify which candidate solutions work.

Second, AI models often lose track of details over multi-day tasks because attention disperses across growing context windows. To prevent this, /teamwork partitions context by construction: each specialized agent receives only the exact files and tools it needs for a specific step. Coordinators hand off tasks to fresh successors, while critics and auditors challenge and verify results from clean contexts without the original producer's reasoning.

Teamwork expresses these workflows as declarative self-improving natural-language patterns and dynamically routes each user prompt to the pattern best suited to the task. These patterns range from distributed coding and iterative problem-solving (packaged standalone as /boost, a sequential solve-and-verify loop refined from prior trajectories) to specialized research workflows like long proof and document review, all of which are available in Antigravity today. This allows teams to design custom specifications from past run transcripts routing across different harness patterns and dynamically improving agentic harness. While many of these features are available in Antigravity, we will provide more updates in the coming weeks providing all the features to our external users on Antigravity. Following our previous posts on Antigravity /teamwork, we provide new results to open challenges empowered by new variants of /teamwork and Gemini 4 models.

Mathematics and theoretical computer science

Across mathematics and theoretical computer science, researchers paired with Gemini 4 and /teamwork harnesses to collaborate on large-scale proofs and generate complete algorithms and proofs directly from open problem statements. We present several examples.

Proving and verifying the Courtade–Kumar conjecture

Researchers from Google, CUHK, USC, and CMU used /teamwork's Long Proof pattern (implementing Stellar Colosseum in Antigravity) to prove information theory's 2013 Courtade–Kumar conjecture in full generality. The result confirms that when a random bit string passes through a noisy channel, no compression of the input to a single bit preserves more information than a "dictator function", which just copies a single input bit. Gemini first found counterexamples to a generalized conjecture from the CUHK team's earlier paper. The CUHK team then proposed a new candidate function that reduced the conjecture to a 14-variable optimization problem, along with a key analytical lower bound. Gemini 4 Argon and earlier Gemini models then partitioned the parameter space and developed most of the new analytical bounds and interval-arithmetic certificates. They also extended the proof from balanced to unbalanced functions without human input. The team formalized the entire proof in Lean 4, producing 50 million lines of code.

Resolving open questions in optimization, statistics, quantum computing, and algorithms

Researchers also posed open problems directly to Gemini 4 and to multi-agent harnesses such as Stellar Colosseum and /teamwork, receiving end-to-end algorithms and proofs (often with Lean 4 formalizations) that the authors verified, simplified, and edited for publication. In adaptive data analysis, a multi-agent harness running on Gemini 4 Argon independently proved and Lean-verified near-optimal sample bounds for statistical queries chosen adaptively from earlier answers, closing a gap open since 2014. These bounds yield the first nearly space-optimal algorithm for data streams with both insertions and deletions that stays accurate against adaptive adversaries. In quantum computing, a new construction settled a 2014 open question by finding the first n-bit Boolean functions needing strictly more than n/2 quantum queries at constant error, while pinning down the exact query overhead across symmetric functions.

Several results break long standing barriers in processing and compressing large matrices and graphs. For fifty years until STOC 2024, combinatorial algorithms for Boolean matrix multiplication and triangle detection (avoiding the algebraic cancellations of fast matrix multiplication) saved only logarithmic factors over cubic time. A new deterministic algorithm resolves STOC 2024's main open question, achieving its target speedup over cubic time. In high-dimensional geometry, a new lower bound resolves an open problem on sparse networks, proving that approximating all pairwise distances among n points within any constant factor can require polynomially more than n edges, even with extra Steiner points and arbitrarily many hops. In randomized linear algebra, stacked sparse sketches resolve a 2025 open question by breaking the classical sparsity barrier for compressing a matrix to linear dimension without collapsing any direction of an r-dimensional subspace, needing only √log r nonzeros per row (quadratically fewer than the log r required to preserve all lengths) and nearly matching a new lower bound. For weighted graphs under dynamic edge-weight insertions and deletions, Stellar Colosseum resolved an open question left by FOCS 2022, developing the first single-pass streaming algorithm for cut sparsification that achieves near-optimal memory, update time, and reconstruction time simultaneously.

Our results also span optimization, arithmetic, data structures, and statistics. In optimization, resolving an open question on online inverse linear optimization (inferring an expert's hidden objective from decisions over time), the approach gave a deterministic algorithm achieving optimal regret while cutting runtime from exponential in the dimension to polynomial time. For n-bit integer multiplication, multi-agent campaigns simplified the theoretical Turing-machine bound below n log n, while a separate effort gave a new O(n log n) algorithm on a single two-dimensional grid whose C implementation matches the GNU Multiple Precision Arithmetic Library (GMP) bit-for-bit up to a billion bits. In data structures, another result made significant progress on Tarjan's 1985 Deque Conjecture (whether self-adjusting splay trees handle insertions and deletions at both ends in constant amortized time), proving a new upper bound on the per-operation overhead that simultaneously beats every finitely iterated inverse-Ackermann function, verified in Lean 4. Finally, in high-dimensional statistics, the approach proved that approximate leave-one-out cross-validation, which avoids refitting the model from scratch, remains accurate for non-smooth sparse penalties like elastic net.

Scaling formal verification and manuscript review

Checking a computer-assisted manuscript by hand places a heavy burden on peer reviewers, and subtle flaws can hide inside long calculations. Formalizing mathematics in Lean 4 turns a proof into code that a computer kernel checks line by line. Once human mathematicians confirm that the formal theorem statement matches the original problem, the proof only compiles if every logical step holds.

Many of the deepest open problems in mathematics govern how physical waves propagate, how dynamical systems evolve, and whether polynomial equations have integer solutions. Working directly from informal proofs, our multi-agent framework, LEAP, developed with colleagues in Google Cloud AI and paired with Argon, has made progress in these core areas.

Notably, /teamwork-LEAP fully autonomously produced an end-to-end, machine-checked formalization of the Weil conjectures, the landmark results on counting solutions of polynomial equations over finite fields that culminated in Pierre Deligne's 1974 proof. This formalization spans about 1.8 million lines of Lean, produced across 28,800 subagent invocations. The system also synthesized over 919,000 lines of Lean to produce the first formalization of a recent proof of Hilbert's tenth problem over general rings of integers. It also independently reproduces an end-to-end formalization of a recent proof of finite-time blowup for the 3D Navier–Stokes equations with smooth forcing, part of one of the seven Millennium Prize Problems and open for about 90 years.

Beyond these, it also generated the first formalization of the Birkhoff conjecture, a dynamical-systems problem open since 1927. Moreover, it formalized for the first time the 3D Bochner–Riesz conjecture, a central problem in harmonic analysis open for more than 50 years, finding and fixing a gap in a recently released informal proof along the way. Additional machine-checked breakthroughs include resolving the C¹ regularity of infinity-harmonic functions in three and higher dimensions (an open question for about 20 years), formalizing Killip and Vişan's theorem that the KdV water-wave equation is well-posed in H⁻¹ and verifying a UCLA group's informal proofs of the 40-year-old Jamison Caterpillar Conjecture, the Ağırseven-Ollis conjecture, and an improved upper bound on Erdős Problem #1063.

Finally, we provided Deep Review for authors of another CS conference: ahead of STOC 2027, authors of 578 submissions (over a third of all submissions) opted in to receive Deep Review feedback on their drafts. In an ongoing author survey (N = 70), 90% of respondents rated the feedback mostly or almost all factual and 90% said they would use the tool again. Over 30% said Deep Review caught substantive proof bugs or mathematically significant typos, and 14% said it found technical gaps that took more than an hour to fix. Surveyed authors shared that “Tools of this kind can help produce cleaner and more precise research, faster” and that DeepReview provides peace of mind on scientific discovery”. We provide users two weekly calls to Deep Review for free. Deep Review is also implemented as /teamwork's document review pattern and available in Antigravity. Read more about DeepReview in our recent blog post.

Molecular chemistry discovery

Beyond formal proofs, pairing specialized generator agents with independent verifier agents enables discovery workflows in experimental laboratory science.

Determining a molecule’s structure from experimental nuclear magnetic resonance data is one of chemistry’s hardest puzzles. For example, in HIV protease inhibitors, one misplaced bond separates an inactive look-alike from a drug. By organizing 30 specialized agents into a self-correcting loop with /teamwork and a physics-informed spectral simulator, we achieved 95% top-1 exact match and 100% top-3 accuracy on complex benchmark molecules, nearly doubling single-agent performance. Beyond teaching Gemini to reason over structure, this lays the groundwork for ongoing works on 3D protein and complex structures dynamics.

Beyond this example, science problems including applications in nuclear fusion and quantum algorithms reported in the Gemini 4 blog post benefited from /teamwork and the Argon model.

Generating a causal inference benchmark

We also applied a similar multi-round auditing workflow to build empirical benchmarks. Starting from five human-specified acceptance criteria like data anonymization, 29 agents running on Argon over 11.5 hours turned 13 public tables into a 50-task causal inference benchmark. In these tasks, models must determine whether a causal effect can be identified and estimate it against a known answer. After initial checks, separate reviewer and auditor agents caught shortcuts, flagged wording hints, and vetoed hallucinated scripts, turning each into a permanent test. Automating this heavily manual task generation supplies the hard problems needed to help Gemini answer counterfactual "what-if" questions.

Coding and systems engineering

When paired with objective evaluators such as network benchmarks, game engines, unmodified test suites, and hardware timers, /teamwork campaigns can execute multi-day software and hardware engineering projects.

Distributed systems

Using /teamwork and Argon, we built FarArray, a low-latency kernel-bypass Partitioned Global Address Space based on the SNAP microkernel. Applications built on FarArray include lock-free Connected Components, which scale to 1-trillion-edge graphs with 60x to 90x throughput gains and 40x to 180x speedups at 6x lower RAM over TCP key-value stores. Similarly, InstantEmbedding (Personalized PageRank) reduced peak DRAM by up to 14x and runtime by 11x to 27x over accelerated MapReduce implementations. The team also used /teamwork and Argon to create custom TPU kernels for Graph Neural Networks, delivering up to a 185x High Bandwidth Memory reduction and 8x to 15x speedups in forward and backward passes. Further details are available in our recent Antigravity blog post.

Autonomous hill climbing

To test autonomous optimization over long horizons, hill-climbing teams on Argon trained a 9x9 Go engine from scratch starting from a basic AlphaZero implementation. Human researchers set the chip allocation and evaluation setup, while the harness ran autonomous campaigns for faster self-play (achieving a 2,300x speedup) and faster Elo readings, promoting champions strictly via statistical tests on locked records. In under eight days, the engine gained roughly 2,600 Elo, reaching 1,885 Elo, placing it between KataGo at 16 and 32 visits per move.

Memory-safe Rust rewrites, /boost, and TPU kernel optimization

Large code migrations fail when agents stub out dependencies or alter tests to make a broken port pass. A new /teamwork rewrite pattern, soon to be available in Antigravity, solved this by scheduling each module only after its dependencies compile, linking every leaf module against the original, unmodified test suite. Building on the initial 25% port of the Zircon microkernel, a two-week Argon campaign rewrote the remaining C++ code, adding about 800,000 lines of safe Rust that pass all kernel unit tests and boot on arm64 hardware. /teamwork also rewrote libgav1, Google's AV1 video decoder, from C++ into around 170k lines Rust, achieving 100% bit-exact parity across a broad conformance test suite. Initially, the port was around 4x slower than the original C++ code. However, after giving /teamwork access to profiling tools and a mandate to maximize speed on sample video clips, the Rust port became 20-30% faster than the original C++ and hand-optimized assembly code on both x86_64 and ARM.

For day-to-day coding tasks, the Antigravity /boost command packages /teamwork's iterative coding pattern. A long-running /teamwork campaign analyzed past run trajectories and found that tightly coupled code is best handled through sequential implementation and revision stages rather than parallel splits, improving /boost on coding tasks. This is an example of a harness optimizing another harness for token efficiency.

For machine learning systems, full hardware evaluations of custom JAX and Pallas TPU kernels take 7 to 32 minutes, limiting how many candidates prior optimization agents can explore. Using /teamwork's DuoDiscover pattern on Argon, four parallel agents identified performance bottlenecks and constructed fast, 23-second proxy evaluators called yardsticks. These yardsticks produce candidate orderings that are continuously validated against periodic full hardware runs using Kendall's rank correlation. DuoDiscover achieves a 4.09x geometric-mean speedup over reference implementations across 17 JAXBench kernels and a 2.99x geometric-mean speedup across 7 hand-optimized vLLM TPU-Inference Pallas kernels for frontier open-weight models.

Google infrastructure applications

We have also applied /teamwork and /boost to many production infrastructure and hardware modeling problems across Google. Examples include:

  • Fleet-wide memory optimization: Building on the initial Argon memory efforts, we analyzed fleet-wide metrics, as well as internal memory saving past successes, and created reusable /teamwork workflows that have identified memory saving opportunities, surpassing our initial savings estimates of 1 PiB. We also deployed the human-in-the-loop Memory Efficiency Monitoring Agent, achieving additional memory savings, using techniques such as struct padding elimination, scalar narrowing, cliff jumping, outlining, and tagged pointers.
  • Borg agentic scheduling: Built on /teamwork and /boost, Borg’s agentic workflow for scheduling optimization shipped 2 successful changes, resulting in a 24% reduction in TPU workload preemption on some hardware types, while preserving workload SLOs and cluster utilization.

Opening these tools to the research and developer community

External scientists and developers can use Gemini 4 models with /teamwork and /boost in Antigravity and Deep Review today. We have also opened applications giving 10,000 academic researchers six months of free access to our highest usage tier. Upcoming updates will bring our formal proof pattern and additional specialized workflows into Antigravity. These features will be accompanied by deep-dive tutorials on the Antigravity blog in the coming weeks, ensuring that developers and researchers have the resources needed to generate, inspect, and iteratively refine custom multi-agent teams. We are excited to make our tools (in addition to our models) available for everyone.

Acknowledgements

Antigravity’s /teamwork feature is a result of a Google-wide effort with many contributors and ongoing explorations. Contributions include: Yuan Deng[279eb3], Honghao Lin[279eb3], Jieming Mao[279eb3], Vahab Mirrokni[279eb3], Vishnu Veeramachaneni[279eb3], David Woodruff[279eb3], Song Zuo[279eb3], Shubham Agarwal, Santiago Balseiro, MohammadHossein Bateni, Mohsen Bayati, Garrett Bingham, Guy Blelloch, Douglas Chen, Vincent Cohen-Addad, Piaoyang Cui, Tulsee Doshi, Eric First, Michael Galkin, Elena Gribelyuk, Anupam Gupta, Guru Guruganesh, Mike Heddes, Patrick Heisel, Iman Hosseini, Kevin Hou, Lalit Jain, Adel Javanmard, Rajesh Jayaram, Xiaoen Ju, Po-Nien Kung, Silvio Lattanzi, Samuel Lavoie, Chun-Liang Li, Matthew Li, Yinan Li, Yuchen Li, Yiming Liu, Thang Luong, Brandon Mayer, Simon Meierhans, Swaroop Mishra, Ashish Naik, Sebastian Nowozin, Renato Paes Leme, Nikos Parotsidis, Ali Parviz, Neel Patel, Violet Peng, Bryan Perozzi, Anshul Ramachandran, Shashir Reddy, Lijie Ren, Madhavi Sewak, D Shin, Reza Shokri, Sukhdeep Singh, Samuel Smith, Linfeng Song, Dilara Soylu, Zheng Tan, Sasan Tavakkol, Hadi Tavassol, Yifeng Teng, Tom Wollschläger, Haifeng Xu, Zicheng Xu, Yihao Xue, Amir Yazdanbakhsh, and Goran Žužić. Carla Bromberg and Lauren Fraser for program management, and Francis Duan, Pan-Pan Jiang, Xiyan Shao, Jane Shapiro, Weilun Shi, and Xinyu Ye for their contributions to the project. We also thank support from several leaders Sergey Brin, Demis Hassabis, Koray Kavukcuoglu, Pushmeet Kohli, James Manyika, Yossi Matias, Varun Mohan, and Amin Vahdat.

Several academic collaborators contributing to the research papers: Zijie Chen, Amin Gohari, Chandra Nair (CUHK), Songbin Liu, Arian Maleki, Chen Shang, Hanbo Yang, Wenjin Zhang (Columbia University), Zihan Hao, Qipeng Liu, Jianan Shao (UCSD), Lorenzo Beretta (IBM Cambridge), Simone Bombari (BYU & Flatiron Institute), Mahbod Majid, Margalit Glasgow (MIT), Raphael Arkady Meyer (UC Berkeley, ICSI, & LBNL), Kijun Shin (CMU), Huacheng Yu, Elena Gribelyuk (Princeton University), Xin Lyu (UCB/ ByteDance), Samson Zhou (Texas A&M), Erik Waingarten (UPenn), and Sorrachai Yingchareonthawornchai (ETH-ITS).

原文版权归 Google Research 及原作者所有,本站转载仅供阅读。如需删除,请联系 support@aisa.one。