{"@context":"https://schema.org","@type":"NewsArticle","generatedAt":"2026-10-07T11:00:56.538Z","headline":"Google 发布 Gemini 4 Argon：缩小与 OpenAI 和 Anthropic 的差距但未明显领先","description":"Google 发布新旗舰模型 Gemini 4 Argon，是继 Gemini 3.1 Pro 之后七个多月来的首款前沿模型。","url":"https://www.aioga.com/news/oyqytutjzet3wan8qz7ovr37m/","mainEntityOfPage":"https://www.aioga.com/news/oyqytutjzet3wan8qz7ovr37m/","datePublished":"2026-09-30T22:06:30.000Z","dateModified":"2026-09-30T22:06:30.000Z","inLanguage":"zh-CN","publisher":{"@type":"NewsMediaOrganization","name":"Aioga","url":"https://www.aioga.com"},"citation":["https://the-decoder.com/google-gemini-4-argon-closes-the-gap-with-openai-and-anthropic-but-doesnt-take-a-clear-lead","https://aihot.news/items/oyqytutjzet3wan8qz7ovr37m"],"canonicalUrl":"https://www.aioga.com/news/oyqytutjzet3wan8qz7ovr37m/","directAnswer":{"@type":"Answer","text":"Aioga 编辑摘要：Google 发布新旗舰模型 Gemini 4 Argon，是继 Gemini 3.1 Pro 之后七个多月来的首款前沿模型。 Aioga 将其归入「行业动态」方向，重点关注它对真实使用和行业竞争的影响。","url":"https://www.aioga.com/news/oyqytutjzet3wan8qz7ovr37m/","dateCreated":"2026-09-30T22:06:30.000Z","author":{"@type":"Organization","@id":"https://www.aioga.com/authors/aioga-editorial/#editorial-team","name":"Aioga Editorial Team","url":"https://www.aioga.com/authors/aioga-editorial/"}},"evidence":[{"@type":"CreativeWork","name":"the-decoder.com source article","url":"https://the-decoder.com/google-gemini-4-argon-closes-the-gap-with-openai-and-anthropic-but-doesnt-take-a-clear-lead","datePublished":"2026-09-30T22:06:30.000Z","provider":{"@type":"Organization","name":"the-decoder.com","url":"https://the-decoder.com/google-gemini-4-argon-closes-the-gap-with-openai-and-anthropic-but-doesnt-take-a-clear-lead"}},{"@type":"CreativeWork","name":"AIHot archive record","url":"https://aihot.news/items/oyqytutjzet3wan8qz7ovr37m","datePublished":"2026-09-30T22:06:30.000Z","provider":{"@type":"Organization","name":"AIHot","url":"https://aihot.news/items/oyqytutjzet3wan8qz7ovr37m"}}],"aggregationSource":"The Decoder：AI News（RSS）","originalPublisher":{"name":"the-decoder.com","url":"https://the-decoder.com/google-gemini-4-argon-closes-the-gap-with-openai-and-anthropic-but-doesnt-take-a-clear-lead"},"geoDeepAnswer":null,"article":{"id":"oyqytutjzet3wan8qz7ovr37m","slug":"oyqytutjzet3wan8qz7ovr37m","url":"https://www.aioga.com/news/oyqytutjzet3wan8qz7ovr37m/","title":"Google 发布 Gemini 4 Argon：缩小与 OpenAI 和 Anthropic 的差距但未明显领先","title_en":"","summary":"Google 发布新旗舰模型 Gemini 4 Argon，是继 Gemini 3.1 Pro 之后七个多月来的首款前沿模型。","source":"The Decoder：AI News（RSS）","sourceUrl":"https://the-decoder.com/google-gemini-4-argon-closes-the-gap-with-openai-and-anthropic-but-doesnt-take-a-clear-lead","aiHotUrl":"https://aihot.news/items/oyqytutjzet3wan8qz7ovr37m","publishedAt":"2026-09-30T22:06:30.000Z","category":"行业动态","score":58,"selected":false,"articleBody":["Google unveiled Gemini 4 Argon, its new frontier model that closes the gap with rivals from OpenAI and Anthropic, beating some of them on key benchmarks. While it may not clearly lead the pack, it is relatively cheap for a frontier model, at least at the introductory price.","Argon is Google's first frontier model in more than seven months, following Gemini 3.1 Pro. It puts the ad giant back among the top three AI labs, though Anthropic likely still holds the lead. After a difficult and drawn-out development period：https://the-decoder.de/googles-ki-krise-deepmind-verliert-unabhaengigkeit-und-hassabis-ist-so-gut-wie-weg/ that saw the already-announced Gemini 3.5 frontier model skipped entirely, Google is back in the race.","Argon is initially going to a group of \"trusted cyber defenders\" as part of the Fairwind program：https://deepmind.google/fairwind-program/. They and Google's internal teams will get the model without cyber guardrails. Google justifies the gradual rollout with a \"phased approach\" that AI capabilities at this level require. The company is also taking part in the US government's voluntary program that gives agencies access to new models before public release. Ad","Pricing is already set, at least as an introductory rate: $2 per million input tokens and $10 per million output tokens. Cached input tokens cost 95 percent less, working out to about 10 cents per million. Gemini 3.8 Flash had a 90 percent cache discount. That puts Google well below other frontier models on raw token price, though not on token consumption (see below).","*Google doesn't state this figure explicitly but says the cache is 95 percent cheaper than the regular input token price. Ad","Google also raised the output limit from 64,000 to one million tokens, calling it an industry first. The idea is that if the model can generate hundreds of thousands of tokens in a single trajectory, it can think through hard problems more thoroughly and solve them in one pass. To support this, Google is adding a new \"Long Decode Continuation\" feature to the Gemini API. It pauses long responses and resumes them through follow-up requests so reasoning doesn't hit a timeout.","The input context window stays at one million tokens. Argon accepts text, images, video, and audio as input but only outputs text. Ad","Artificial Analysis：https://artificialanalysis.ai/models/gemini-4-argon provides an early independent assessment. At its highest available reasoning level, \"High,\" Gemini 4 Argon scores 53 points on the Artificial Analysis Intelligence Index. That ties it with OpenAI's GPT-6 Astra (max) and Claude Fable 5.1, and puts it one point ahead of GPT-6.1 Sol (max). Ad","Anthropic's models still lead. Claude Opus 5.5 sits at 58 points and Claude Sonnet 5.5 at 56. Compared to Google's last frontier model, Gemini 3.1 Pro Preview, Argon jumped 23 points. \"High\" is typically the top reasoning tier for Gemini models, though a special \"Deep Think\" mode is sometimes supported as well.","At the current promo price, one Intelligence Index task costs $1.99. That's 60 percent of GPT-6 Astra's cost ($3.26) but 2.7 times more expensive than GPT-6.1 Sol. Once the discount ends, the cost rises to $3.98, about 20 percent above GPT-6 Astra. The price advantage comes from lower token rates, not from efficiency. Argon uses an average of 62,000 output tokens per task, while GPT-6 Astra needs only 27,000.","Argon also made significant gains on agentic tasks, which according to Artificial Analysis have been a weak spot for Gemini models. On AutomationBench-AA, the Artificial Analysis variant, it takes first place at 77.5 percent, six points ahead of Claude Sonnet 5.5 (max). On Terminal Bench 4, it hits 57 percent, a 53-point jump over Gemini 3.1 Pro Preview. That still leaves it behind Claude Sonnet 5.5 (64 percent), Claude Opus 5.5 (60 percent), and GPT-6 Astra (59 percent).","Artificial Analysis also highlights Argon's low hallucination rate. On AA-Omniscience, a benchmark that tests factual knowledge and honest handling of knowledge gaps, Argon's hallucination rate is 15 percent. GPT-6 Astra (max) comes in at 51 percent and GPT-6.1 Sol (max) at 54 percent. Argon is far more likely to admit it doesn't know an answer rather than guess wrong. Its accuracy, however, reaches only 50 percent, five points below Gemini 3.1 Pro Preview and 13 points below GPT-6 Astra (max, 63 percent). On the benchmark's overall score, Argon lands at 42 points, roughly even with GPT-6 Astra (43) and GPT-6.1 Sol (42).","Google's own benchmark results：https://storage.googleapis.com/deepmind-media/gemini/gemini_4_argon_model_evaluation.pdf paint a rosier picture. Argon leads in most of those benchmarks, sometimes by wide margins.","Argon also leads the Vals Index：https://x.com/ValsAI/status/2105388446844072033. At 68.9 percent, it takes first place according to Vals AI, making it the first Gemini model to top the index. Argon finishes in the top five on 20 of 22 tested benchmarks, with particular strength in finance, law, coding, and security. In this test, though, the mid-tier model Sonnet 5.5：https://the-decoder.com/anthropics-claude-sonnet-5-5-nearly-matches-opus-5-5-on-benchmarks-while-costing-up-to-30-percent-less-per-task/ also outranks Anthropic's top model Opus 5.5, so take it with a grain of salt.","As always, AI models have to prove themselves in real-world use, and performance depends not just on the model itself but also on the software wrapped around it. That's Google's weak spot right now: compared to Claude Cowork and ChatGPT Work, the Gemini app still lags behind.","On Arena.ai：https://x.com/arena/status/2105394855644139908, where humans rate model outputs in head-to-head comparisons, Argon performs well. In the Text Arena, Gemini 4 Argon (High) takes first place with 1,525 points, 20 points ahead of Claude Opus 4.6 (High) in second. That makes it a strong contender for writing tasks, especially after a long drought of competitive writing models and Opus 5.5 still not matching Opus 4.6：https://the-decoder.com/anthropic-engineer-explains-why-claudes-writing-got-worse-although-the-model-got-smarter/ in human preference rankings. Google's previous model, Gemini 3.8 Flash (High), had been in eleventh place.","According to Arena, Argon leads in coding, hard prompts, instruction following, longer queries, and creative writing. It also ranks first across all evaluated professional fields, as well as for queries in English, Chinese, Russian, and non-English queries overall.","Web development results are more modest. In Code Arena: WebDev, Argon scores 1,679 points and lands in eighth place. That's a 96-point improvement over Gemini 3.8 Flash (High) and a jump from 29th, but it doesn't crack the top spots.","On price-to-performance, Arena puts Argon ahead of the field. At a blended rate of $8 per million tokens, Argon shifts the Pareto frontier of the Text Arena and is currently the most cost-efficient model in the ranking.","Stay in the loop on AI. Clear, useful, no fluff.","Follow The Decoder for AI news, background stories and expert analyses.","The Decoder：https://the-decoder.com/"],"articleImages":[{"sourceUrl":"https://the-decoder.com/wp-content/uploads/2026/09/gemini_4_argon.png","alt":"Image description","afterParagraph":0,"url":"/media/articles/oyqytutjzet3wan8qz7ovr37m/20f57e728bfe5fbe.png"},{"sourceUrl":"https://the-decoder.com/wp-content/uploads/2026/09/AA_gemini_4_high.jpg","alt":"","afterParagraph":9,"url":"/media/articles/oyqytutjzet3wan8qz7ovr37m/c3eb4cbf76a4e549.jpg"},{"sourceUrl":"https://the-decoder.com/wp-content/uploads/2026/09/gemini-4-argon_table_blog-scaled-1.gif","alt":"","afterParagraph":12,"url":"/media/articles/oyqytutjzet3wan8qz7ovr37m/c04c84a2139c8517.gif"}],"mediaStatus":"ok","articleBodyZh":["谷歌发布了Gemini 4 Argon，这是其新的前沿模型，缩小了与OpenAI和Anthropic竞争对手的差距，并在一些关键基准测试中超过了它们中的一些。虽然它可能没有明显领先，但对于前沿模型来说，它的价格相对便宜，至少在介绍价格阶段如此。","Argon是谷歌在七个多月以来推出的首个前沿模型，继Gemini 3.1 Pro之后。它使这家广告巨头重新进入前三大AI实验室的行列，尽管Anthropic很可能仍然占据领先地位。在经历了一个艰难且漫长的开发周期之后：https://the-decoder.de/googles-ki-krise-deepmind-verliert-unabhaengigkeit-und-hassabis-ist-so-gut-wie-weg/ 已经宣布的Gemini 3.5前沿模型被完全跳过，谷歌重新回到了竞争中。","Argon最初将分发给一批“受信任的网络防御者”，作为Fairwind计划的一部分：https://deepmind.google/fairwind-program/。他们以及谷歌的内部团队将获得不带网络防护措施的模型。谷歌以AI在这个水平上需要“分阶段方法”来证明逐步推出的合理性。该公司还参与了美国政府的自愿计划，使各机构在模型公开发布前获得访问权限。广告","定价已经确定，至少作为初始费率：输入令牌每百万$2，输出令牌每百万$10。缓存的输入令牌成本低95%，折合约每百万$0.10。Gemini 3.8 Flash有90%的缓存折扣。这使得谷歌在原始令牌价格上远低于其他前沿模型，但在令牌消耗方面则不然（见下文）。","*谷歌没有明确说明这个数字，但表示缓存比常规输入令牌价格便宜95%。广告","谷歌还将输出限制从64,000令牌提高到一百万令牌，并称之为行业首例。其理念是，如果模型能在单次轨迹中生成数十万个令牌，它就可以更彻底地思考复杂问题，并一次性解决它们。为支持这一点，谷歌在Gemini API中新增了“长解码续延”(Long Decode Continuation)功能。它可以暂停长响应，并通过后续请求恢复，以防推理超时。","输入上下文窗口保持在一百万令牌。Argon可以接受文本、图像、视频和音频作为输入，但只输出文本。广告","人工分析：https://artificialanalysis.ai/models/gemini-4-argon 提供了早期的独立评估。在其最高可用推理水平“High”下，Gemini 4 Argon 在人工分析智能指数中得分为53分。这与 OpenAI 的 GPT-6 Astra（最高）和 Claude Fable 5.1 持平，领先 GPT-6.1 Sol（最高）一分。","Anthropic 的模型仍然领先。Claude Opus 5.5 得分58分，Claude Sonnet 5.5 得分56分。相比 Google 的最后一代模型 Gemini 3.1 Pro Preview，Argon 提升了23分。“High” 通常是 Gemini 模型的顶级推理层次，尽管有时也支持特殊的“Deep Think”模式。","以当前促销价格计算，一个智能指数任务费用为1.99美元，仅为 GPT-6 Astra（3.26美元）成本的60%，但比 GPT-6.1 Sol 高 2.7 倍。一旦折扣结束，费用将上升至3.98美元，比 GPT-6 Astra 高约20%。价格优势来自较低的 token 费率，而非效率。Argon 每个任务平均使用62,000个输出 token，而 GPT-6 Astra 仅需27,000个。","Argon 在代理任务上也取得了显著进展，而根据人工分析，这一直是 Gemini 模型的弱项。在 AutomationBench-AA（人工分析版本）中，Argon 以77.5%的成绩位居第一，比 Claude Sonnet 5.5（最高）高出六分。在 Terminal Bench 4 上，它达到57%，比 Gemini 3.1 Pro Preview 提升53分。但仍落后于 Claude Sonnet 5.5（64%）、Claude Opus 5.5（60%）和 GPT-6 Astra（59%）。","人工分析还强调了 Argon 的低幻觉率。在 AA-Omniscience 上，这是一个测试事实知识及对知识空缺诚实处理能力的基准测试，Argon 的幻觉率为15%。GPT-6 Astra（最高）为51%，GPT-6.1 Sol（最高）为54%。Argon 更倾向于承认自己不知道答案，而不是猜错。然而，其准确率仅为50%，比 Gemini 3.1 Pro Preview 低五分，比 GPT-6 Astra（最高，63%）低13分。在该基准测试的总体评分中，Argon 获得42分，约与 GPT-6 Astra（43）和 GPT-6.1 Sol（42）持平。","谷歌自有的基准测试结果：https://storage.googleapis.com/deepmind-media/gemini/gemini_4_argon_model_evaluation.pdf 描绘了更乐观的情况。Argon 在大多数基准测试中领先，有时优势很大。","Argon 也在 Vals 指数中领先：https://x.com/ValsAI/status/2105388446844072033。以 68.9% 的得分，根据 Vals AI，它名列第一，成为首个登顶该指数的 Gemini 模型。Argon 在 22 项测试的基准中，有 20 项进入前五，尤其在金融、法律、编程和安全领域表现强劲。不过，在这项测试中，中端模型 Sonnet 5.5：https://the-decoder.com/anthropics-claude-sonnet-5-5-nearly-matches-opus-5-5-on-benchmarks-while-costing-up-to-30-percent-less-per-task/ 也超过了 Anthropic 的顶级模型 Opus 5.5，因此结果需谨慎看待。","一如既往，AI 模型必须在实际使用中证明自身，性能不仅取决于模型本身，还取决于其周围的软件。这正是谷歌目前的弱点：与 Claude Cowork 和 ChatGPT Work 相比，Gemini 应用仍然落后。","在 Arena.ai：https://x.com/arena/status/2105394855644139908 上，人类通过一对一比较对模型输出进行评分，Argon 表现良好。在 Text Arena 中，Gemini 4 Argon (High) 以 1,525 分名列第一，领先第二名 Claude Opus 4.6 (High) 20 分。这使其成为写作任务的有力竞争者，尤其在长时间没有竞争性写作模型的情况下，而 Opus 5.5 仍未在人工偏好排名中超过 Opus 4.6：https://the-decoder.com/anthropic-engineer-explains-why-claudes-writing-got-worse-although-the-model-got-smarter/。谷歌之前的模型 Gemini 3.8 Flash (High) 曾排在第十一位。","根据 Arena 的数据，Argon 在编程、困难提示、遵循指令、长查询和创意写作方面领先。它在所有评估的专业领域，以及英语、中文、俄语和非英语查询中均排名第一。","网页开发的结果较为平淡。在 Code Arena: WebDev 中，Argon 得分 1,679 分，位居第八。相比 Gemini 3.8 Flash (High)，提升了 96 分，从第 29 位跃升，但仍未进入前列。","在性价比方面，Arena 将 Argon 放在了领先位置。以每百万令牌 8 美元的综合价格计算，Argon 改变了 Text Arena 的帕累托前沿，目前是排行榜中成本效率最高的模型。","保持对 AI 的关注。清晰、有用，无冗余。","关注 The Decoder，获取 AI 新闻、背景故事和专家分析。","解码器：https://the-decoder.com/"],"translationStatus":"translated","bodyOrigin":"source-page","editorial":{"summary":"Aioga 编辑摘要：Google 发布新旗舰模型 Gemini 4 Argon，是继 Gemini 3.1 Pro 之后七个多月来的首款前沿模型。 Aioga 将其归入「行业动态」方向，重点关注它对真实使用和行业竞争的影响。","background":"背景分析：公司与行业类动态需要放在竞争格局、商业化路径、资本信号和监管环境中观察，单条公告不能代表最终结果。","viewpoint":"Aioga 判断：这条动态更适合作为行业观察信号，当前信息足以建立线索，但不足以推导长期结论。","implications":"影响分析：对相关团队而言，短期应先核对来源、可用范围和实际成本，再判断是否值得接入或跟进。","nextStep":"后续观察：继续观察官方文件、合作落地、收入或用户信号、竞品动作和监管后续。","evidenceRefs":["title","summary","articleBody"],"confidence":"medium","status":"published","aiGenerated":false,"autoApproved":true,"generatedBy":"rule-safe-fallback","generatedAt":"2026-10-07T11:11:26.242Z","sourceHash":"2085a4f9ce89799b","validation":{"passed":true,"mode":"rule-safe-fallback","checks":["schema","length","source-attribution","no-html"]}},"tags":["行业动态","The Decoder：AI News（RSS）"],"translations":{"zh-CN":{"title":"Google 发布 Gemini 4 Argon：缩小与 OpenAI 和 Anthropic 的差距但未明显领先","summary":"Google 发布新旗舰模型 Gemini 4 Argon，是继 Gemini 3.1 Pro 之后七个多月来的首款前沿模型。","category":"行业动态","source":"the-decoder.com","aggregationSource":"The Decoder：AI News（RSS）","pageTitle":"Google 发布 Gemini 4 Argon：缩小与 OpenAI 和 Anthropic 的差距但未明显领先 - Aioga AI资讯","description":"Google 发布新旗舰模型 Gemini 4 Argon，是继 Gemini 3.1 Pro 之后七个多月来的首款前沿模型。","url":"https://www.aioga.com/news/oyqytutjzet3wan8qz7ovr37m/","articleBody":["谷歌发布了Gemini 4 Argon，这是其新的前沿模型，缩小了与OpenAI和Anthropic竞争对手的差距，并在一些关键基准测试中超过了它们中的一些。虽然它可能没有明显领先，但对于前沿模型来说，它的价格相对便宜，至少在介绍价格阶段如此。","Argon是谷歌在七个多月以来推出的首个前沿模型，继Gemini 3.1 Pro之后。它使这家广告巨头重新进入前三大AI实验室的行列，尽管Anthropic很可能仍然占据领先地位。在经历了一个艰难且漫长的开发周期之后：https://the-decoder.de/googles-ki-krise-deepmind-verliert-unabhaengigkeit-und-hassabis-ist-so-gut-wie-weg/ 已经宣布的Gemini 3.5前沿模型被完全跳过，谷歌重新回到了竞争中。","Argon最初将分发给一批“受信任的网络防御者”，作为Fairwind计划的一部分：https://deepmind.google/fairwind-program/。他们以及谷歌的内部团队将获得不带网络防护措施的模型。谷歌以AI在这个水平上需要“分阶段方法”来证明逐步推出的合理性。该公司还参与了美国政府的自愿计划，使各机构在模型公开发布前获得访问权限。广告","定价已经确定，至少作为初始费率：输入令牌每百万$2，输出令牌每百万$10。缓存的输入令牌成本低95%，折合约每百万$0.10。Gemini 3.8 Flash有90%的缓存折扣。这使得谷歌在原始令牌价格上远低于其他前沿模型，但在令牌消耗方面则不然（见下文）。","*谷歌没有明确说明这个数字，但表示缓存比常规输入令牌价格便宜95%。广告","谷歌还将输出限制从64,000令牌提高到一百万令牌，并称之为行业首例。其理念是，如果模型能在单次轨迹中生成数十万个令牌，它就可以更彻底地思考复杂问题，并一次性解决它们。为支持这一点，谷歌在Gemini API中新增了“长解码续延”(Long Decode Continuation)功能。它可以暂停长响应，并通过后续请求恢复，以防推理超时。","输入上下文窗口保持在一百万令牌。Argon可以接受文本、图像、视频和音频作为输入，但只输出文本。广告","人工分析：https://artificialanalysis.ai/models/gemini-4-argon 提供了早期的独立评估。在其最高可用推理水平“High”下，Gemini 4 Argon 在人工分析智能指数中得分为53分。这与 OpenAI 的 GPT-6 Astra（最高）和 Claude Fable 5.1 持平，领先 GPT-6.1 Sol（最高）一分。","Anthropic 的模型仍然领先。Claude Opus 5.5 得分58分，Claude Sonnet 5.5 得分56分。相比 Google 的最后一代模型 Gemini 3.1 Pro Preview，Argon 提升了23分。“High” 通常是 Gemini 模型的顶级推理层次，尽管有时也支持特殊的“Deep Think”模式。","以当前促销价格计算，一个智能指数任务费用为1.99美元，仅为 GPT-6 Astra（3.26美元）成本的60%，但比 GPT-6.1 Sol 高 2.7 倍。一旦折扣结束，费用将上升至3.98美元，比 GPT-6 Astra 高约20%。价格优势来自较低的 token 费率，而非效率。Argon 每个任务平均使用62,000个输出 token，而 GPT-6 Astra 仅需27,000个。","Argon 在代理任务上也取得了显著进展，而根据人工分析，这一直是 Gemini 模型的弱项。在 AutomationBench-AA（人工分析版本）中，Argon 以77.5%的成绩位居第一，比 Claude Sonnet 5.5（最高）高出六分。在 Terminal Bench 4 上，它达到57%，比 Gemini 3.1 Pro Preview 提升53分。但仍落后于 Claude Sonnet 5.5（64%）、Claude Opus 5.5（60%）和 GPT-6 Astra（59%）。","人工分析还强调了 Argon 的低幻觉率。在 AA-Omniscience 上，这是一个测试事实知识及对知识空缺诚实处理能力的基准测试，Argon 的幻觉率为15%。GPT-6 Astra（最高）为51%，GPT-6.1 Sol（最高）为54%。Argon 更倾向于承认自己不知道答案，而不是猜错。然而，其准确率仅为50%，比 Gemini 3.1 Pro Preview 低五分，比 GPT-6 Astra（最高，63%）低13分。在该基准测试的总体评分中，Argon 获得42分，约与 GPT-6 Astra（43）和 GPT-6.1 Sol（42）持平。","谷歌自有的基准测试结果：https://storage.googleapis.com/deepmind-media/gemini/gemini_4_argon_model_evaluation.pdf 描绘了更乐观的情况。Argon 在大多数基准测试中领先，有时优势很大。","Argon 也在 Vals 指数中领先：https://x.com/ValsAI/status/2105388446844072033。以 68.9% 的得分，根据 Vals AI，它名列第一，成为首个登顶该指数的 Gemini 模型。Argon 在 22 项测试的基准中，有 20 项进入前五，尤其在金融、法律、编程和安全领域表现强劲。不过，在这项测试中，中端模型 Sonnet 5.5：https://the-decoder.com/anthropics-claude-sonnet-5-5-nearly-matches-opus-5-5-on-benchmarks-while-costing-up-to-30-percent-less-per-task/ 也超过了 Anthropic 的顶级模型 Opus 5.5，因此结果需谨慎看待。","一如既往，AI 模型必须在实际使用中证明自身，性能不仅取决于模型本身，还取决于其周围的软件。这正是谷歌目前的弱点：与 Claude Cowork 和 ChatGPT Work 相比，Gemini 应用仍然落后。","在 Arena.ai：https://x.com/arena/status/2105394855644139908 上，人类通过一对一比较对模型输出进行评分，Argon 表现良好。在 Text Arena 中，Gemini 4 Argon (High) 以 1,525 分名列第一，领先第二名 Claude Opus 4.6 (High) 20 分。这使其成为写作任务的有力竞争者，尤其在长时间没有竞争性写作模型的情况下，而 Opus 5.5 仍未在人工偏好排名中超过 Opus 4.6：https://the-decoder.com/anthropic-engineer-explains-why-claudes-writing-got-worse-although-the-model-got-smarter/。谷歌之前的模型 Gemini 3.8 Flash (High) 曾排在第十一位。","根据 Arena 的数据，Argon 在编程、困难提示、遵循指令、长查询和创意写作方面领先。它在所有评估的专业领域，以及英语、中文、俄语和非英语查询中均排名第一。","网页开发的结果较为平淡。在 Code Arena: WebDev 中，Argon 得分 1,679 分，位居第八。相比 Gemini 3.8 Flash (High)，提升了 96 分，从第 29 位跃升，但仍未进入前列。","在性价比方面，Arena 将 Argon 放在了领先位置。以每百万令牌 8 美元的综合价格计算，Argon 改变了 Text Arena 的帕累托前沿，目前是排行榜中成本效率最高的模型。","保持对 AI 的关注。清晰、有用，无冗余。","关注 The Decoder，获取 AI 新闻、背景故事和专家分析。","解码器：https://the-decoder.com/"]},"en":{"title":"Google Releases Gemini 4 Argon: Narrows the Gap with OpenAI and Anthropic but Does Not Clearly Take the Lead","summary":"Google releases its new flagship model, Gemini 4 Argon, its first frontier model in more than seven months since Gemini 3.1 Pro.","category":"Industry","source":"The Decoder：AI News（RSS）","aggregationSource":"The Decoder：AI News（RSS）","pageTitle":"Google Releases Gemini 4 Argon: Narrows the Gap with OpenAI and Anthropic but Does Not Clearly Take the Lead - Aioga AI News","description":"Google releases its new flagship model, Gemini 4 Argon, its first frontier model in more than seven months since Gemini 3.1 Pro.","url":"https://www.aioga.com/en/news/oyqytutjzet3wan8qz7ovr37m/","contentTranslated":true,"translationStatus":"translated","translationRetryAt":"","translationError":"","sourceHash":"d0f93d0116cf85fb","translatedAt":"2026-10-01T03:39:38.502Z"},"ja":{"title":"Google、Gemini 4 Argonを発表：OpenAIやAnthropicとの差を縮めるも、明確な優位には立たず","summary":"Googleは新しいフラッグシップモデルGemini 4 Argonを発表しました。これはGemini 3.1 Pro以来、7か月以上ぶりの最初の先端モデルです。","category":"業界動向","source":"The Decoder：AI News（RSS）","aggregationSource":"The Decoder：AI News（RSS）","pageTitle":"Google、Gemini 4 Argonを発表：OpenAIやAnthropicとの差を縮めるも、明確な優位には立たず - Aioga AIニュース","description":"Googleは新しいフラッグシップモデルGemini 4 Argonを発表しました。これはGemini 3.1 Pro以来、7か月以上ぶりの最初の先端モデルです。","url":"https://www.aioga.com/ja/news/oyqytutjzet3wan8qz7ovr37m/","contentTranslated":true,"translationStatus":"translated","translationRetryAt":"","translationError":"","sourceHash":"d0f93d0116cf85fb","translatedAt":"2026-10-01T03:40:00.330Z"},"ko":{"title":"Google, Gemini 4 Argon 출시: OpenAI 및 Anthropic과의 격차를 좁혔지만 뚜렷한 우위는 확보하지 못해","summary":"Google이 새로운 플래그십 모델 Gemini 4 Argon을 출시했다. 이는 Gemini 3.1 Pro 이후 7개월여 만에 나온 첫 번째 최첨단 모델이다.","category":"업계 동향","source":"The Decoder：AI News（RSS）","aggregationSource":"The Decoder：AI News（RSS）","pageTitle":"Google, Gemini 4 Argon 출시: OpenAI 및 Anthropic과의 격차를 좁혔지만 뚜렷한 우위는 확보하지 못해 - Aioga AI 뉴스","description":"Google이 새로운 플래그십 모델 Gemini 4 Argon을 출시했다. 이는 Gemini 3.1 Pro 이후 7개월여 만에 나온 첫 번째 최첨단 모델이다.","url":"https://www.aioga.com/ko/news/oyqytutjzet3wan8qz7ovr37m/","contentTranslated":true,"translationStatus":"translated","translationRetryAt":"","translationError":"","sourceHash":"d0f93d0116cf85fb","translatedAt":"2026-10-01T03:39:26.283Z"},"es":{"title":"Google lanza Gemini 4 Argon: reduce la brecha con OpenAI y Anthropic, pero no logra una ventaja clara","summary":"Google presenta su nuevo modelo insignia, Gemini 4 Argon, el primer modelo de vanguardia en más de siete meses desde Gemini 3.1 Pro.","category":"Industria","source":"The Decoder：AI News（RSS）","aggregationSource":"The Decoder：AI News（RSS）","pageTitle":"Google lanza Gemini 4 Argon: reduce la brecha con OpenAI y Anthropic, pero no logra una ventaja clara - Aioga Noticias de IA","description":"Google presenta su nuevo modelo insignia, Gemini 4 Argon, el primer modelo de vanguardia en más de siete meses desde Gemini 3.1 Pro.","url":"https://www.aioga.com/es/news/oyqytutjzet3wan8qz7ovr37m/","contentTranslated":true,"translationStatus":"translated","translationRetryAt":"","translationError":"","sourceHash":"d0f93d0116cf85fb","translatedAt":"2026-10-01T03:44:38.768Z"},"fr":{"title":"Google lance Gemini 4 Argon : réduit l’écart avec OpenAI et Anthropic, sans toutefois prendre une avance nette","summary":"Google lance son nouveau modèle phare Gemini 4 Argon, le premier modèle de pointe depuis Gemini 3.1 Pro il y a plus de sept mois.","category":"Industrie","source":"The Decoder：AI News（RSS）","aggregationSource":"The Decoder：AI News（RSS）","pageTitle":"Google lance Gemini 4 Argon : réduit l’écart avec OpenAI et Anthropic, sans toutefois prendre une avance nette - Aioga Actualités IA","description":"Google lance son nouveau modèle phare Gemini 4 Argon, le premier modèle de pointe depuis Gemini 3.1 Pro il y a plus de sept mois.","url":"https://www.aioga.com/fr/news/oyqytutjzet3wan8qz7ovr37m/","contentTranslated":true,"translationStatus":"translated","translationRetryAt":"","translationError":"","sourceHash":"d0f93d0116cf85fb","translatedAt":"2026-10-01T03:44:09.236Z"},"de":{"title":"Google veröffentlicht Gemini 4 Argon: Verringert den Abstand zu OpenAI und Anthropic, liegt aber nicht deutlich vorn","summary":"Google hat das neue Flaggschiffmodell Gemini 4 Argon veröffentlicht, das erste Spitzenmodell seit Gemini 3.1 Pro nach mehr als sieben Monaten.","category":"行业动态","source":"The Decoder：AI News（RSS）","aggregationSource":"The Decoder：AI News（RSS）","pageTitle":"Google veröffentlicht Gemini 4 Argon: Verringert den Abstand zu OpenAI und Anthropic, liegt aber nicht deutlich vorn - Aioga KI-News","description":"Google hat das neue Flaggschiffmodell Gemini 4 Argon veröffentlicht, das erste Spitzenmodell seit Gemini 3.1 Pro nach mehr als sieben Monaten.","url":"https://www.aioga.com/de/news/oyqytutjzet3wan8qz7ovr37m/","contentTranslated":true,"translationStatus":"translated","translationRetryAt":"","translationError":"","sourceHash":"d0f93d0116cf85fb","translatedAt":"2026-10-01T03:44:50.607Z"},"pt-BR":{"title":"Google lança o Gemini 4 Argon: reduz a distância em relação à OpenAI e à Anthropic, mas não assume uma liderança clara","summary":"O Google lançou o novo modelo principal Gemini 4 Argon, sendo o primeiro modelo de ponta em mais de sete meses desde o Gemini 3.1 Pro.","category":"行业动态","source":"The Decoder：AI News（RSS）","aggregationSource":"The Decoder：AI News（RSS）","pageTitle":"Google lança o Gemini 4 Argon: reduz a distância em relação à OpenAI e à Anthropic, mas não assume uma liderança clara - Aioga Notícias de IA","description":"O Google lançou o novo modelo principal Gemini 4 Argon, sendo o primeiro modelo de ponta em mais de sete meses desde o Gemini 3.1 Pro.","url":"https://www.aioga.com/pt-BR/news/oyqytutjzet3wan8qz7ovr37m/","contentTranslated":true,"translationStatus":"translated","translationRetryAt":"","translationError":"","sourceHash":"d0f93d0116cf85fb","translatedAt":"2026-10-01T03:49:16.896Z"},"ru":{"title":"Google представила Gemini 4 Argon: сократила отставание от OpenAI и Anthropic, но не стала явным лидером","summary":"Google выпустила новую флагманскую модель Gemini 4 Argon — первую передовую модель более чем за семь месяцев после Gemini 3.1 Pro.","category":"行业动态","source":"The Decoder：AI News（RSS）","aggregationSource":"The Decoder：AI News（RSS）","pageTitle":"Google представила Gemini 4 Argon: сократила отставание от OpenAI и Anthropic, но не стала явным лидером - Aioga Новости ИИ","description":"Google выпустила новую флагманскую модель Gemini 4 Argon — первую передовую модель более чем за семь месяцев после Gemini 3.1 Pro.","url":"https://www.aioga.com/ru/news/oyqytutjzet3wan8qz7ovr37m/","contentTranslated":true,"translationStatus":"translated","translationRetryAt":"","translationError":"","sourceHash":"d0f93d0116cf85fb","translatedAt":"2026-10-01T03:48:41.196Z"},"ar":{"title":"تُطلق Google نموذج Gemini 4 Argon: تُضيّق الفجوة مع OpenAI وAnthropic، لكنها لا تتقدم عليهما بشكل واضح","summary":"أعلنت Google عن نموذجها الرائد الجديد Gemini 4 Argon، وهو أول نموذج متقدم تطلقه منذ أكثر من سبعة أشهر بعد Gemini 3.1 Pro.","category":"行业动态","source":"The Decoder：AI News（RSS）","aggregationSource":"The Decoder：AI News（RSS）","pageTitle":"تُطلق Google نموذج Gemini 4 Argon: تُضيّق الفجوة مع OpenAI وAnthropic، لكنها لا تتقدم عليهما بشكل واضح - Aioga أخبار الذكاء الاصطناعي","description":"أعلنت Google عن نموذجها الرائد الجديد Gemini 4 Argon، وهو أول نموذج متقدم تطلقه منذ أكثر من سبعة أشهر بعد Gemini 3.1 Pro.","url":"https://www.aioga.com/ar/news/oyqytutjzet3wan8qz7ovr37m/","contentTranslated":true,"translationStatus":"translated","translationRetryAt":"","translationError":"","sourceHash":"d0f93d0116cf85fb","translatedAt":"2026-10-01T03:48:15.167Z"},"hi":{"title":"Google ने Gemini 4 Argon जारी किया: OpenAI और Anthropic के साथ अंतर कम हुआ, लेकिन स्पष्ट बढ़त नहीं मिली","summary":"Google ने नया फ्लैगशिप मॉडल Gemini 4 Argon जारी किया है, जो Gemini 3.1 Pro के बाद सात महीने से अधिक समय में जारी किया गया पहला फ्रंटियर मॉडल है।","category":"行业动态","source":"The Decoder：AI News（RSS）","aggregationSource":"The Decoder：AI News（RSS）","pageTitle":"Google ने Gemini 4 Argon जारी किया: OpenAI और Anthropic के साथ अंतर कम हुआ, लेकिन स्पष्ट बढ़त नहीं मिली - Aioga AI समाचार","description":"Google ने नया फ्लैगशिप मॉडल Gemini 4 Argon जारी किया है, जो Gemini 3.1 Pro के बाद सात महीने से अधिक समय में जारी किया गया पहला फ्रंटियर मॉडल है।","url":"https://www.aioga.com/hi/news/oyqytutjzet3wan8qz7ovr37m/","contentTranslated":true,"translationStatus":"translated","translationRetryAt":"","translationError":"","sourceHash":"d0f93d0116cf85fb","translatedAt":"2026-10-01T03:53:31.690Z"},"it":{"title":"Google ha rilasciato Gemini 4 Argon: riduce il divario con OpenAI e Anthropic ma non prende chiaramente il sopravvento","summary":"Google ha lanciato il nuovo modello di punta Gemini 4 Argon, il primo modello all'avanguardia in oltre sette mesi dopo il Gemini 3.1 Pro.","category":"行业动态","source":"The Decoder：AI News（RSS）","aggregationSource":"The Decoder：AI News（RSS）","pageTitle":"Google ha rilasciato Gemini 4 Argon: riduce il divario con OpenAI e Anthropic ma non prende chiaramente il sopravvento - Aioga Notizie IA","description":"Google ha lanciato il nuovo modello di punta Gemini 4 Argon, il primo modello all'avanguardia in oltre sette mesi dopo il Gemini 3.1 Pro.","url":"https://www.aioga.com/it/news/oyqytutjzet3wan8qz7ovr37m/","contentTranslated":true,"translationStatus":"translated","translationRetryAt":"","translationError":"","sourceHash":"d0f93d0116cf85fb","translatedAt":"2026-10-01T03:52:59.632Z"},"nl":{"title":"Google brengt Gemini 4 Argon uit: verkleint de kloof met OpenAI en Anthropic, maar neemt geen duidelijke voorsprong","summary":"Google heeft het nieuwe vlaggenschipmodel Gemini 4 Argon uitgebracht, het eerste frontiermodel in meer dan zeven maanden sinds Gemini 3.1 Pro.","category":"行业动态","source":"The Decoder：AI News（RSS）","aggregationSource":"The Decoder：AI News（RSS）","pageTitle":"Google brengt Gemini 4 Argon uit: verkleint de kloof met OpenAI en Anthropic, maar neemt geen duidelijke voorsprong - Aioga AI-nieuws","description":"Google heeft het nieuwe vlaggenschipmodel Gemini 4 Argon uitgebracht, het eerste frontiermodel in meer dan zeven maanden sinds Gemini 3.1 Pro.","url":"https://www.aioga.com/nl/news/oyqytutjzet3wan8qz7ovr37m/","contentTranslated":true,"translationStatus":"translated","translationRetryAt":"","translationError":"","sourceHash":"d0f93d0116cf85fb","translatedAt":"2026-10-01T03:53:17.311Z"},"tr":{"title":"Google, Gemini 4 Argon'u Yayınladı: OpenAI ve Anthropic ile Arayı Kısaltıyor ama Belirgin Bir Şekilde Önde Değil","summary":"Google, Gemini 3.1 Pro’dan yedi aydan uzun bir süre sonra ilk öncü modeli olan yeni amiral gemisi modeli Gemini 4 Argon’u piyasaya sürdü.","category":"行业动态","source":"The Decoder：AI News（RSS）","aggregationSource":"The Decoder：AI News（RSS）","pageTitle":"Google, Gemini 4 Argon'u Yayınladı: OpenAI ve Anthropic ile Arayı Kısaltıyor ama Belirgin Bir Şekilde Önde Değil - Aioga AI Haberleri","description":"Google, Gemini 3.1 Pro’dan yedi aydan uzun bir süre sonra ilk öncü modeli olan yeni amiral gemisi modeli Gemini 4 Argon’u piyasaya sürdü.","url":"https://www.aioga.com/tr/news/oyqytutjzet3wan8qz7ovr37m/","contentTranslated":true,"translationStatus":"translated","translationRetryAt":"","translationError":"","sourceHash":"d0f93d0116cf85fb","translatedAt":"2026-10-01T03:57:36.250Z"},"vi":{"title":"Google ra mắt Gemini 4 Argon: thu hẹp khoảng cách với OpenAI và Anthropic nhưng chưa vượt trội rõ rệt","summary":"Google ra mắt mô hình hàng đầu mới Gemini 4 Argon, mô hình tiên phong đầu tiên sau hơn bảy tháng kể từ Gemini 3.1 Pro.","category":"行业动态","source":"The Decoder：AI News（RSS）","aggregationSource":"The Decoder：AI News（RSS）","pageTitle":"Google ra mắt Gemini 4 Argon: thu hẹp khoảng cách với OpenAI và Anthropic nhưng chưa vượt trội rõ rệt - Tin tức AI Aioga","description":"Google ra mắt mô hình hàng đầu mới Gemini 4 Argon, mô hình tiên phong đầu tiên sau hơn bảy tháng kể từ Gemini 3.1 Pro.","url":"https://www.aioga.com/vi/news/oyqytutjzet3wan8qz7ovr37m/","contentTranslated":true,"translationStatus":"translated","translationRetryAt":"","translationError":"","sourceHash":"d0f93d0116cf85fb","translatedAt":"2026-10-01T03:57:45.497Z"},"id":{"title":"Google merilis Gemini 4 Argon: memperkecil kesenjangan dengan OpenAI dan Anthropic tetapi belum jelas memimpin","summary":"Google merilis model flagship baru Gemini 4 Argon, model terdepan pertamanya dalam lebih dari tujuh bulan sejak Gemini 3.1 Pro.","category":"行业动态","source":"The Decoder：AI News（RSS）","aggregationSource":"The Decoder：AI News（RSS）","pageTitle":"Google merilis Gemini 4 Argon: memperkecil kesenjangan dengan OpenAI dan Anthropic tetapi belum jelas memimpin - Berita AI Aioga","description":"Google merilis model flagship baru Gemini 4 Argon, model terdepan pertamanya dalam lebih dari tujuh bulan sejak Gemini 3.1 Pro.","url":"https://www.aioga.com/id/news/oyqytutjzet3wan8qz7ovr37m/","contentTranslated":true,"translationStatus":"translated","translationRetryAt":"","translationError":"","sourceHash":"d0f93d0116cf85fb","translatedAt":"2026-10-01T03:58:05.279Z"},"th":{"title":"Google เปิดตัว Gemini 4 Argon: ลดช่องว่างกับ OpenAI และ Anthropic แต่ยังไม่ได้เหนือกว่าอย่างชัดเจน","summary":"Google เปิดตัวโมเดลเรือธงรุ่นใหม่ Gemini 4 Argon ซึ่งเป็นโมเดลแนวหน้ารุ่นแรกในรอบกว่าเจ็ดเดือนนับตั้งแต่ Gemini 3.1 Pro","category":"行业动态","source":"The Decoder：AI News（RSS）","aggregationSource":"The Decoder：AI News（RSS）","pageTitle":"Google เปิดตัว Gemini 4 Argon: ลดช่องว่างกับ OpenAI และ Anthropic แต่ยังไม่ได้เหนือกว่าอย่างชัดเจน - ข่าว AI Aioga","description":"Google เปิดตัวโมเดลเรือธงรุ่นใหม่ Gemini 4 Argon ซึ่งเป็นโมเดลแนวหน้ารุ่นแรกในรอบกว่าเจ็ดเดือนนับตั้งแต่ Gemini 3.1 Pro","url":"https://www.aioga.com/th/news/oyqytutjzet3wan8qz7ovr37m/","contentTranslated":true,"translationStatus":"translated","translationRetryAt":"","translationError":"","sourceHash":"d0f93d0116cf85fb","translatedAt":"2026-10-01T04:02:17.889Z"},"pl":{"title":"Google wydało Gemini 4 Argon: zmniejsza dystans do OpenAI i Anthropic, ale nie osiąga wyraźnej przewagi","summary":"Google zaprezentowało nowy flagowy model Gemini 4 Argon, pierwszy model z najwyższej półki od ponad siedmiu miesięcy, od czasu Gemini 3.1 Pro.","category":"行业动态","source":"The Decoder：AI News（RSS）","aggregationSource":"The Decoder：AI News（RSS）","pageTitle":"Google wydało Gemini 4 Argon: zmniejsza dystans do OpenAI i Anthropic, ale nie osiąga wyraźnej przewagi - Aioga Wiadomości AI","description":"Google zaprezentowało nowy flagowy model Gemini 4 Argon, pierwszy model z najwyższej półki od ponad siedmiu miesięcy, od czasu Gemini 3.1 Pro.","url":"https://www.aioga.com/pl/news/oyqytutjzet3wan8qz7ovr37m/","contentTranslated":true,"translationStatus":"translated","translationRetryAt":"","translationError":"","sourceHash":"d0f93d0116cf85fb","translatedAt":"2026-10-01T04:02:16.081Z"}},"evidenceTier":"verified-news","reviewStatus":"automated-ingest","indexable":true,"editorialCover":""}}