（文/陈济深 编辑/张广凯） 12月10日，彭博社爆料称，扎克伯格组成了一个名为TBD Lab的团队。该团队在其最新模型"牛油 果"（Avocado）的训练过程中使用了多个第三方模型，包括谷歌的Gemma、OpenAI的GPT-oss和阿里巴 巴的Qwen模型。该款模型预计将于明年春季首次亮相，并可能作为"闭源"模型推出。 对此，Meta 的一位发言人则对外宣称："我们的模型训练工作正按计划进行，时间表没有发生有意义的 变更。" 消息被爆出后，阿里巴巴美股盘前一度上涨4%，收盘涨幅2.53%。 扎克伯格挥舞重金招来的AI大牛们计划开发的闭源大模型，竟然是通过中国的开源模型来训练，这不 仅意味着如今中国开源阵营的崛起，也代表扎克伯格曾经的美国开源霸主豪言，终究没有抵过来自中国 的竞争压力。 过气的开源盟主 Meta急着抄作业源于其开源旗舰模型Llama 4的失败表现。 过去两年，Meta通过开源Llama系列，成功扮演了"反OpenAI联盟"的盟主。Llama被视为开源界的 Linux，一度是全球开发者（包括中国开发者）的首选底座。 然而，这一格局在2025年开始瓦解。 随着年初DeepSeek开源模型的横空出 ...

人民财讯12月11日电，12月11日，智谱正式发布工业级语音合成系统GLM-TTS，并在Hugging Face和 ModelScope上开放模型权重。GLM-TTS目前已开源，并上线智谱开放平台(bigmodel)、智谱清言、 Z.ai。 ...

【导读】 扎克伯格使用阿里千问优化Meta新AI模型 中国基金报记者 泰勒 大家好，简单关注一下国产AI大模型的一则利好。 12月10日晚间，阿里巴巴美股盘前突然直线拉升，正式开盘后，维持2%左右的涨幅。 | 最高:160.74 | 今开:159.64 | 成交量:414.83万股 | 换手:0.17% | | --- | --- | --- | --- | | 最低:158.41 | 昨收:155.96 | 成交额:6.61亿 | 振幅:1.49% | | 52周最高: 192.67 | 量比:6.28 | 市盈率(TTM): 21.81 | 市净率:2.61 | | 52周最低: 79.43 | 委比: 0.00% | 市盈率(静): 20.67 | 市销率:2.50 | | 每股收益:7.32 | 股息(TTM): 2.00 | 每手股数:1 | 总市值:3808.25亿 | | 每股净资产:7.66 | 股息率(TTM): 1.28% | 最小价差:0.01 | 总股本:23.86亿 | | 机构持股: -- | Beta: -- | 空头回补天数:一 | 货币单位:USD | | | AND THE ...

据知情人士透露，一款代号为"Avocado"的新模型预计将在明年春天亮相，且很可能会以"闭源"模型的 形式推出——一种可以被严格控制、并由Meta对外出售使用权的模型。这一做法与竞争对手谷歌和 OpenAI的路线一致，将成为与Meta多年来大力宣扬的开源战略迄今为止最大的一次背离。开源模型允 许外部开发者和研究人员审查并在代码基础上进行二次开发。据这些知情人士说，Meta新任首席AI官 Alexandr Wang倾向于支持闭源模型。 今年早些时候，在Meta发布了开源模型Llama 4之后，公司的战略出现了重大转向。Llama 4的表现让硅 谷和扎克伯格都感到失望。扎克伯格把部分参与该项目的人边缘化，同时亲自出面挖角顶尖的AI研究 人员和负责人，有些人拿到了数亿美元级别、为期多年的薪酬方案，现在，扎克伯格把大量时间和精力 花在和这些新加入的人密切合作上，他们被集中到一个名为TBD Lab的团队中。 据知情人士透露，TBD Lab在训练Avocado的过程中，正在使用多家第三方模型，从包括谷歌的 Gemma、OpenAI的gpt-oss以及阿里巴巴集团旗下的通义千问（Qwen）等竞争对手模型中进行"蒸馏"。 用 ...

【导读】 扎克伯格使用阿里千问优化Meta新AI模型 中国基金报记者 泰勒 大家好，简单关注一下国产AI大模型的一则利好。 12月10日晚间，阿里巴巴美股盘前突然直线拉升，正式开盘后，维持2%左右的涨幅。 | 最高: 160.74 | 今开:159.64 | | --- | --- | | 最低:158.41 | 昨收:155.96 | | 52周最高: 192.67 | 量比: 6.28 | | 52周最低: 79.43 | 委比:0.00% | | 每股收益:7.32 | 股息(TTM): 2.00 | | 每股净资产:7.66 | 股息率(TTM): 1.28% | | 机构持股: -- | Beta: -- | 编辑：杜妍 校对：纪元 审核：陈思扬 | 量: 414.83万股 | 换手: 0.17% | | --- | --- | | 额:6.61亿 | 振幅:1.49% | | 率(TTM): 21.81 | 市净率:2.61 | | 率(静): 20.67 | 市销率:2.50 | | 般数:1 | 总市值:3808.25亿 | | 价差:0.01 | 总股本:23.86亿 | | 回补天数: - ...

知情人士表示，TBD Lab在Avocado模型的训练过程中，引入了多款第三方模型的技术成果，对包括谷 歌的Gemma模型、OpenAI的gpt-oss，以及阿里巴巴的通义千问（Qwen）模型在内的竞品技术进行了 提炼整合。 Meta的发言人对此事拒绝置评。 据知情人士透露，一款代号为Avocado的全新模型预计将于明年春季正式推出，且大概率会以闭源模式 面市 —— 这种模式下模型可被严格管控，Meta能够通过授权使用的方式进行销售。此举与竞争对手谷 歌（Google）和OpenAI的模型运营策略一致，也标志着Meta彻底背离其多年来标榜的开源战略。开源 模型允许外部开发者与研究人员查看代码，并在此基础上进行二次开发。知情人士补充称，Meta新任 人工智能首席官Alexandr Wang是闭源模型的坚定支持者。 今年早些时候，Meta推出开源模型Llama 4，但这款产品的表现令硅谷业界及扎克伯格本人均倍感失 望，公司的人工智能战略也由此发生重大转向。扎克伯格不仅将该项目的部分核心成员调离核心岗位， 还亲自出面招揽顶尖人工智能研究人才与团队负责人，为部分人选开出了数亿美元的多年期薪酬方案。 其中，Alexan ...

彭博社报道称，扎克伯格在组建科技史上最昂贵的团队之一数月后，开始亲自参与日常工作，并将公司的重心从开源模型转向可以盈利的闭源AI模型。Meta明年将推出代号“Avocado（牛油果）”的闭源新模型，与OpenAI展开竞争。Meta新任首席人工智能官 Alexandr Wang（汪滔）是封闭式模型的拥护者。 https://t.co/PLy4HjZIXC ...

加州大学伯克利分校（UC Berkeley）刚刚发布了一份重磅论文：《Measuring Agents in Production》。 （论文地址：https://arxiv.org/pdf/2512.04123） 这份论文，基于来自全球的真实请求：306名从业者深度调研，20个企业级部署案例，覆盖 26 个行业。 这是AI Agent 领域，迄今最大规模的实证研究。 最核心的三个信息： 这份报告信息非常多，容我慢慢道来。 73%为生产力买单，金融成Agent 第一战场 先说一个数字： 73%的从业者表示，部署Agent的首要目的是"提高生产力"。 其中，金融与银行业是第一大战场，占比39.1% 其次是科技（24.6%）和企业服务（23.2%） 。 除了这些，Agent 还在很多意想不到的地方落地： 保险理赔流程自动化：代理人负责处理从保单查询到风险识别的序列排序流程。 生物医学工作流自动化：在科学发现领域，Agent 用于自动化执行复杂的实验和数据分析流程。 企业内部运营支持：涵盖人力资源信息搜索、站点故障事件诊断等多个方面。 这些跨行业的成功案例证明，AI Agent已经具备解决真实世界复杂问题的能力，并 ...

Core Insights - The report titled "State of AI: An Empirical 100 Trillion Token Study with OpenRouter" analyzes the usage of over 300 models on the OpenRouter platform from November 2024 to November 2025, highlighting significant trends in AI development and the growing importance of open-source models [3][5]. Group 1: Open Source vs. Closed Source Models - Open-source models are expected to reach approximately one-third of total usage by the end of the year, complementing rather than competing with closed-source models [5][7]. - The share of Chinese open-source models surged from 1.2% to 30% in weekly usage, indicating a strong preference for domestic models [9]. - The dominance of DeepSeek as the largest contributor is diminishing as more open-source models enter the market, leading to a more diversified landscape by mid-2025 [12]. Group 2: Model Characteristics and Trends - The report categorizes models into large (700 billion parameters or more), medium (150 to 700 billion), and small (less than 150 billion), noting a shift towards medium and large models as small models lose popularity [15]. - Language models are evolving from dialogue systems to reasoning and execution systems, with reasoning token usage exceeding 50% [18][19]. - The use of tools within models is increasing, indicating a more competitive and diverse ecosystem [24]. Group 3: Usage Patterns and User Retention - AI model usage has shifted from simple tasks to more complex problem-solving, with average input prompts increasing fourfold [26][30]. - The concept of "Cinderella effect" describes how users may quickly adopt new models that perfectly meet their needs, leading to high retention rates for successful models [57][59]. - Programming and role-playing are now the primary use cases for AI models, with programming queries rising from 11% to over 50% [36][40]. Group 4: Market Dynamics and Regional Insights - The paid usage of AI in Asia has doubled from 13% to 31%, while North America's market share has fallen below 50% [61]. - English remains the dominant language in AI usage at 82%, with Simplified Chinese holding a 5% share [61]. - The impact of model pricing on usage is minimal, with a 10% price drop resulting in only a 0.5%-0.7% increase in usage [61].

Core Insights - The report titled "State of AI: An Empirical 100 Trillion Token Study with OpenRouter" analyzes the usage of over 300 AI models on the OpenRouter platform from November 2024 to November 2025, focusing on real token consumption rather than benchmark scores [3][5][67] - It highlights the significant rise of open-source models, particularly from China, which saw weekly token usage share increase from 1.2% to 30%, indicating a shift towards a complementary relationship between open-source and closed-source models [2][10][74] - The report emphasizes the transition of AI models from language generation systems to reasoning and execution systems, with reasoning models becoming the new paradigm [18][83] Open-Source vs Closed-Source Models - Open-source models are no longer seen merely as alternatives to closed-source models; they have carved out unique positions and are often preferred in specific scenarios [6][70] - By the end of 2025, it is expected that open-source models will account for approximately one-third of total usage, reflecting a more integrated approach by developers who utilize both types of models [5][70] - The dominance of DeepSeek is diminishing as more open-source models enter the market, leading to a diversified landscape where no single model is expected to exceed 25% of token usage by the end of 2025 [13][77] Model Characteristics and Trends - The report identifies a shift towards medium-sized models, which are gaining market favor, while small models are losing traction [16][80] - The classification of models is as follows: large models (700 billion parameters or more), medium models (150 to 700 billion parameters), and small models (less than 150 billion parameters) [20][85] - The usage of reasoning tokens has surpassed 50%, indicating a significant evolution in how AI models are utilized for complex tasks [18][83] User Behavior and Model Utilization - AI model usage has evolved from simple tasks to more complex problem-solving, with user prompts increasing in length and complexity [27][92] - The concept of "crystal shoe effect" is introduced, where certain models lock in a core user base due to their unique capabilities, making it difficult for competitors to attract these users later [55][120] - Programming and role-playing have emerged as the primary use cases for AI models, with programming queries rising from 11% to over 50% [27][100] Market Dynamics - The report notes that the paid usage share of AI in Asia has doubled from 13% to 31%, while North America's share has fallen below 50% [129] - English remains the dominant language in AI usage at 82%, with Simplified Chinese holding nearly 5% [129] - The impact of model pricing on usage is less significant than anticipated, with a 10% price drop leading to only a 0.5%-0.7% increase in usage [129]