跳到正文
11月19日周四
  1. IT之家82

    谷歌 Genie 4 世界模型开放体验,Project Genie 新增智能体与画廊

    谷歌面向符合资格的 Google AI Ultra 订阅用户推出全新 Project Genie 体验,首次开放 Genie 4 访问权限。Genie 4 支持在会话中途输入提示词实时修改已创建的世界,可切换天气、添加角色、改变动作走向,并首次为视觉内容搭配实时原生音频。

    同一件事,精选收录的是《Google DeepMind 发布 Genie 4 并升级 Project Genie》

  2. Rohan Paul85

    Google 发布新模型 Gemini 4 Neon,在 Artificial Analysis 智能指数上得 61 分,排名第三,定价为每 100 万输入/输出 token 3/15 美元。

    引用Sundar Pichai@sundarpichai

    Gemini 4 Neon: frontier-level intelligence with fast end-to-end task completion across complex software engineering, knowledge work, and science workflows. Begins rolling out today in the Gemini API in @GoogleAIStudio, Gemini for business, for paid Google AI Pro and Ultra subscribers in Google @Antigravity, and for Google AI Pro and Ultra subscribers in the @GeminiApp.

    同一件事,精选收录的是《Google 发布 Gemini 4,上线 Gemini App、Antigravity 与 API》

  3. OpenRouter78

    Google DeepMind 的 Gemini 4 Neon 已在 OpenRouter 上线,面向复杂软件工程、知识工作和科学场景,定位介于 Carbon 与 Hydrogen 之间。据 Google 报告,该模型在 Artificial Analysis Coding Agent Index 上排名第一,单任务成本比 Carbon 低 32%。

    引用Models on OpenRouter@ModelsOnOR

    Gemini 4 Neon from @GoogleDeepMind is now live on OpenRouter, via @googlecloud. Built for complex software engineering, knowledge work, and science. It sits between Carbon and Hydrogen. Google reports #1 on the Artificial Analysis Coding Agent Index and a 32% lower cost per task than Carbon.

    推荐理由:Gemini 4 Neon 在 OpenRouter 上线,定位介于 Carbon 与 Hydrogen 之间,并给出编码智能体指数与单任务成本对比。

  4. Andrew Curran67

    Big news: Gemini 4 Neon (High) is the best @GoogleDeepmind model on the Agent Arena Index, ranking #2 overall and reshaping the Pareto frontier! With 1336 pts based on 4.2k real-world agentic sessions from Arena’s global community, it surpassed: - #4 Carbon (High): 1303 pts - #9 Hydrogen (High): 1271 pts - #11 Argon (High): 1229 pts Neon is on the Agent Arena Pareto frontier at $2.82 median cost per task. It scores higher than #3 Claude Opus 5.5 (Max) for 53% less per task, and is 69% less per task than #1 Claude Fable 5.2 (Max). See Pareto in thread. It's strong across key signals: - #2 Praise vs. Complaint: 1539 pts (user satisfaction with the agent’s outcome) - #2 Steerability: 1298 pts (how well an agent responds to user corrections) - #3 Task Completion: 1316 pts (explicit confirmation of task success from users) See where it landed in the Code Arena: WebDev and Text Arena below in thread. Congrats to the team at @GoogleDeepmind on another huge release!

    引用Arena.ai@arena

    Big news: Gemini 4 Neon (High) is the best @GoogleDeepmind model on the Agent Arena Index, ranking #2 overall and reshaping the Pareto frontier! With 1336 pts based on 4.2k real-world agentic sessions from Arena’s global community, it surpassed: - #4 Carbon (High): 1303 pts - #9 Hydrogen (High): 1271 pts - #11 Argon (High): 1229 pts Neon is on the Agent Arena Pareto frontier at $2.82 median cost per task. It scores higher than #3 Claude Opus 5.5 (Max) for 53% less per task, and is 69% less per task than #1 Claude Fable 5.2 (Max). See Pareto in thread. It's strong across key signals: - #2 Praise vs. Complaint: 1539 pts (user satisfaction with the agent’s outcome) - #2 Steerability: 1298 pts (how well an agent responds to user corrections) - #3 Task Completion: 1316 pts (explicit confirmation of task success from users) See where it landed in the Code Arena: WebDev and Text Arena below in thread. Congrats to the team at @GoogleDeepmind on another huge release!

    推荐理由:提供了Gemini 4 Neon在Agent Arena的具体排名(#2)、详细得分(1336 pts)、成本优势($2.82 median cost)及与其他模型(如Claude Opus 5.5, Fable 5.2)的对比数据,这些具体评测结果是此前未披露的新增信息。

  5. Andrew Curran58

    Vals AI 称 Google DeepMind 的 Gemini 4 Neon 在 Vals Index 上以 70.9% 排名第二,比 Claude Fable 5.2 低 0.4 分,单任务成本不到后者一半。它还在 Finance、Legal Index 以及 Tax Agent、Legal Research 上排名第一。榜单图片显示 Gemini 4 Neon 准确率 70.87%、单次测试成本 $7.40、延迟 31m 25s,Claude Fable 5.2 为 71.31%、$17.94、43m 42s。

    引用Vals AI@ValsAI

    Google DeepMind's Gemini 4 Neon debuts at #2 on the Vals Index (70.9%), 0.4 points behind Claude Fable 5.2 at less than half the cost per task. It also takes #1 on our Finance and Legal Indexes, Tax Agent and Legal Research.

  6. Google DeepMind88

    Google DeepMind 发布 Gemini 4 Neon 模型

    Google DeepMind 发布 Gemini 4 Neon,主打复杂软件工程、知识工作与科学流程的快速端到端任务完成。

    推荐理由:官方给出 Neon 在 Gemini 4 家族中的定位、定价与多项基准排名,便于判断其与 Carbon、Hydrogen 的分工。

  7. The Decoder82

    Google 发布 Gemini 4 Neon,一个月内第四款 Gemini 新模型

    Google 发布 Gemini 4 Neon,称其具备前沿级智能,可快速端到端完成复杂软件工程、知识工作和科学工作流任务。该模型今日起通过 API 提供,定价为每百万输入 token 3 美元、每百万输出 token 15 美元,高于 Carbon 和 Argon 的 2/10 美元首发价,也是 Hydrogen 1.50/7.50 美元的两倍。

    同一件事,精选收录的是《Google DeepMind 发布 Gemini 4 Neon 模型》

  8. Demis Hassabis75

    Demis Hassabis 在 X 上给出博客链接,指向 Gemini 4 Neon 的可用性与更多细节,链接地址为 blog.google 上介绍该模型的页面。推文本身未展开具体能力或参数,相关信息需查看博客原文。

    推荐理由:新报道引入了全新的具体产品 'Gemini 4 Neon' 及其可用性信息,这是此前已公开精选(仅涉及 Carbon、Argon、3.7 Flash)中完全未覆盖的新事实。

  9. 🚨 AI News | TestingCatalog82

    Google 正式发布前沿模型 Gemini 4 Neon,并称其在多项基准上超过 Opus 5.5 和 GPT-6.2 Sol。该模型现已向 Google AI Pro 和 Ultra 订阅用户开放,覆盖 Gemini、Antigravity 和 API。作者引用的 Google 方介绍称,Neon 在智能、速度与可用性之间取得平衡,并已在 Google AI Studio 和 API 上线。

    引用Logan Kilpatrick@OfficialLoganK

    Introducing Gemini 4 Neon! Pushing the frontier across domains and punching above its model class in a meaningful way. To me, Neon feels like the right balance of intelligence, speed, and usability. Available now across @GoogleAIStudio, the API, and more!

    同一件事,精选收录的是《Google 发布 Gemini 4,上线 Gemini App、Antigravity 与 API》

  10. Google DeepMind78

    Google DeepMind 发布 Genie 4 并升级 Project Genie

    Google DeepMind 向符合条件的 Google AI Ultra 订阅用户推出新版 Project Genie,首次开放 Genie 4,可在探索过程中输入新提示词实时修改环境,并首次搭配随场景变化的实时原生音频。

    推荐理由:Genie 4 支持会话中实时改世界并加入原生音频,读者可据此判断交互式世界模型的当前能力边界。

  11. 🚨 AI News | TestingCatalog80

    Gemini 4 Neon 已上架 GCP Vertex AI 模型库,可在 Vertex 全球端点调用,并在 Google 模型目录中标记为 GA,定价为每 100 万 token 输入 $3、输出 $15。Google 将其定位为快速推理模型,在编码、推理和生产智能体任务上的质量接近 Gemini 4 Carbon,速度与成本处于 Flash 档位,并基于 Gemini 4 Hydrogen 提供更深推理。该模型尚未出现在 Gemini API(AI Studio)的模型列表中。

    引用AI Tracker Bot@aitrackerbot

    Gemini 4 Neon is live on Vertex AI 🚀 • gemini-4-neon answers on Vertex's global endpoint and is listed as GA in Google's model catalog • $3 input / $15 output per 1M tokens • Not in the Gemini API (AI Studio) model list yet

    推荐理由:Gemini 4 Neon 在 Vertex AI 上架,并给出定价与定位,读者可据此判断它和 Gemini 4 Carbon、Hydrogen 的档位差异。

  12. Demis Hassabis78

    DeepMind 发布 Genie 4 交互式世界模型,整合实时物理、推理与音频能力,Ultra 订阅用户可在 Project Genie 中试用。Demis Hassabis 表示该模型支持模拟中的模拟,更多信息见官方博客。

    引用Demis Hassabis@demishassabis

    Very excited to share Genie 4, our incredible interactive world model, bringing together real-time physics, reasoning, and audio! Simulations within simulations, makes you wonder just how deep the rabbit hole goes... I have a feeling we will be finding out soon enough...

    同一件事,精选收录的是《Google DeepMind 发布 Genie 4 并升级 Project Genie》

  13. The Decoder88

    Vals AI 测试发现 Claude、GPT 与 Gemini 易将平民标为无人机打击目标

    Vals AI 测试发现,Anthropic、OpenAI 和 Google 的模型在无人机打击场景中大多会把平民、救护车、医护人员、记者和投降士兵标记为军事目标,且没有一次被上游安全过滤拦截。

    推荐理由:测试显示多家模型在无人机打击场景中把平民与医护人员标为目标,游戏化措辞即可绕过拒绝。

11月18日周三
  1. Turing Post (Ksenia Se)68

    斯坦福一项论文让 OpenAI、Anthropic、Google 和 Qwen 的 8 个模型回答 620 个真实临床问题,再匿名互评,发现每个评委都偏好自己的答案,平均排名高出 1.22 位,GPT-5.6 Sol 接近 3 位。作者指出,医疗 AI 基准常用 LLM 给 LLM 打分,评委偏好自身输出会让基准偏向该模型;即使去掉评分细则,这种偏好依然存在,公开模型名称也只让偏差缩小 0.094 位。作者建议使用跨模型家族的评委小组,并按评委分别报告结果。

    引用Rohan Paul@rohanpaul_ai

    New Stanford paper finds that all 8 LLMs tested rank their own medical answers higher than other models rank them, even when answers are anonymous. Medical AI benchmarks often use LLMs to grade other LLMs' answers. If a judge favors its own outputs, the benchmark tilts toward that model. They had 8 models from OpenAI, Anthropic, Google and Qwen answer 620 real clinician questions, then rank all 8 anonymous answers. They compared where each judge placed its own answer with where the other 7 judges placed that same answer. Every judge favored its own answer, by 1.22 rank positions on average and nearly 3 for GPT-5.6 Sol. The bias held without a rubric, and revealing model names cut it by only 0.094 positions. If you use LLM judges, use a panel from several model families and report results per judge instead of 1 pooled score.

  2. Orange AI58

    引用信息显示,此前在 Gemini 应用中出现的 Gemini 4 Neon 已在 Google Cloud 计费目录中有了价格,输入 $3、输出 $15 每 1M tokens,Batch 和 Flex 半价为 $1.50/$7.50,支持文本、图像、视频和音频输入,但 Vertex 上尚未开放回答。转发者 Orange AI 认为 Gemini 4 Neon 即将到来,猜测它是 Hydrogen 的另一个 RL 版本,并称其跑分目前是这批里最强的。

    引用AI Tracker Bot@aitrackerbot

    Gemini 4 Neon, the model we caught in the Gemini app on Monday, now has prices in Google Cloud's billing catalog 🔥 • $3 in / $15 out per 1M tokens, live as of 3 AM ET • That puts it between Hydrogen ($1.50 / $7.50) and Carbon ($4 / $20) • Batch and Flex at half price: $1.50 / $7.50 • Takes text, image, video and audio Not answering on Vertex yet. We're watching it closely.