{"id":94896,"topic":"ai","source":"环球网","title":"对标GPT与Claude，谷歌新大模型暂不面向普通用户 - 环球网","url":"https://tech.huanqiu.com/article/4TSPbwKRYkK","url_hash":"a47bc66d67c7f536a28d23daa37ef6b301dd7bd1","author":"","summary":"<a href=\"https://news.google.com/rss/articles/CBMiV0FVX3lxTFBVMVpZLWo1b3JmcVlWSjdyY3RvZ3k1OFdqNW4tU2Qza3dIRmkxX212Yy12R1FJLUEtbEFiNjB5R2xHWnllblZTT1lWLVYxWkJYM043RDhYTQ?oc=5\" target=\"_blank\">对标GPT与Claude，谷歌新大模型暂不面向普通用户</a>&nbsp;&nbsp;<font color=\"#6f6f6f\">环球网</font>","content":"4TSPbwKRYkK tech.huanqiu.comarticle对标GPT与Claude，谷歌新大模型暂不面向普通用户/e3pmh164r/e3pmtmdvg<article><section data-type=\"rtext\"><p>【环球网科技综合报道】10月3日，据外媒CNBC消息，谷歌正式发布新一代前沿旗舰大模型Gemini 4 Argon，直面OpenAI、Anthropic两大对手的竞争。外媒指出，谷歌选择采取受控分阶段的上线策略，暂不对普通用户开放，优先面向网络安全领域经过筛选的专业人员。</p><p><i class=\"pic-con\"><img data-alt=\"\" src=\"//img.huanqiucdn.cn/dp/api/files/imageDir/f1c9d4e4b736f5a9f5eb2309285492ffu1.png?imageView2/2/w/1260\" /></i></p><p>据介绍，Gemini 4 Argon重点强化长周期任务推理、软件工程、企业知识处理与网络安全防御能力，输出Token上限由上代6.4万提升至100万，能够一次性处理数十万Token的复杂任务，已经在谷歌内部用于大型代码库迁移、科研研发等实际工作。在官方19项基准测试中，该模型拿下13项领先成绩。</p> <adv-loader __attr__inner=\"7004636\" __attr__style=\"width: auto;position: relative;float: left;border: 1px solid #ebebeb; padding: 20px;overflow: hidden;margin: 10px 30px 40px 0;\"></adv-loader> <p>在上线节奏上，该模型通过Fairwind项目，仅向经过严格审核的网络安全从业者开放，同时参与美国政府的预发布安全评估。推广期API定价为每百万输入Token2美元、输出 Token10美元，缓存输入可享受大幅折扣，后续将上调价格。面向普通开发者与消费者的全面开放时间暂未公布。</p><p>外媒报道称，近期OpenAI、Anthropic接连曝出前沿大模型安全隐患，AI安全成为行业焦点。分析师表示，谷歌采取保守受控发布模式，意在塑造安全可信的AI厂商定位。但业内也保持理性，基准跑分亮眼不等于实际业务场景同样强大。（旺旺）</p></section></article>1790989252039环球网版权作品，未经书面授权，严禁转载或镜像，违者将被追究法律责任。责编：王楠环球网179098925203911[]{\"email\":\"wangnan@huanqiu.com\",\"name\":\"王楠\"}","image_url":"https://rs2.huanqiucdn.cn/huanqiu/image/www/common/200.jpg","lang":"zh","published_at":"2026-10-02T10:00:00+00:00","fetched_at":"2026-10-03T01:15:07+00:00","status":"read","starred":0,"extract_state":"ok","summary_auto":"4TSPbwKRYkK tech.huanqiu.comarticle对标GPT与Claude，谷歌新大模型暂不面向普通用户/e3pmh164r/e3pmtmdvg<article><section data-type=\"rtext\"><p>【环球网科技综合报道】10月3日，据外媒CNBC消息，谷歌正式发布新一代前沿旗舰大模型Gemini 4 Argon，直面OpenAI、Anthropic两大对手的竞争。外媒指出，谷歌选择采取受控分阶段的上线策略，暂不对普通用户开放，优先面向网络安全领域经过筛选的专业人员。</p><p><i class=\"pic-con\"><img data-alt=\"\" src=\"//img.huanqiucdn.cn/dp/api/files/imageDir/f1c9d4e4b736f5a9f5eb2309285492ffu1.png?imageView2/2/w/1260\" /></i></p><p>据介绍，Gemini 4 Argon重点强化长周期任务推理、软件工程、企业知识处理与网络安全防御能力，输出Token上限由上代6.4万提升至100万，能够一次性处理数十万Token的复杂任务，已经在谷歌内部用于大型代码库迁移、科研研发等实际工作。在官方19项基准测试中，该模型拿下13项领先成绩。</p> <adv-loader __attr__inner=\"7004636\" __attr__style=\"width: auto;position: relative;float: left;border: 1px solid #ebebeb; padding: 20px;overflow: hidden;margin: 10px 30px 40px 0;\"></adv-loader> <p>在上线节奏上，该模型通过Fairwind项目，仅向经过严格审核的网络安全从业者开放，同时参与美国政府的预发布安全评估。推广期API定价为每百万输入Token2美元、输出 Token10美元，缓存输入可享受大幅折扣，后续将上调价格。面向普通开发者与消费者的全面开放时间暂未公布。</p><p>外媒报道称，近期OpenAI、Anthropic接连曝出前沿大模型安全隐患，AI安全成为行业焦点。分析师表示，谷歌采取保守受控发布模式，意在塑造安全可信的AI厂商定位。但业内也保持理性，基准跑分亮眼不等于实际业务场景同样强大。（旺旺）</p></section></article>1790989252039环球网版权作品，未经书面授权，严禁转载或镜像，违者将被追究法律责任。责编：王楠环球网179098925203911[]{\"email\":\"wangnan@huanqiu.com\",\"name\":\"王楠\"}","cluster_id":null,"extract_retries":0,"extract_error":null,"contract_version":"news_item.v1","format_contract_version":"news_item_formats.v1","dedup_url":"https://tech.huanqiu.com/article/4TSPbwKRYkK","quality_profile":{"profile_version":"extraction_quality.v2","bucket":"high","confidence":0.9,"failure_kind":"none","retryable":false,"retry_after_attempts":0,"reason":"High confidence: full text extraction produced 1170 characters.","operator_guidance":{"severity":"ok","recommended_action":"trust_full_text","next_step":"Use the extracted full text as the primary article source.","operator_label":"Ready","can_retry":false,"can_use_summary":false,"diagnostics_required":false},"content_depth":{"contract_version":"content_depth.v1","category":"full_text","label":"Full text","has_full_text":true,"has_summary":true,"content_length":1170,"summary_length":1170,"usable_text_length":1170,"source_field":"content"},"legacy_collapsed":false,"signals":{"extract_state":"ok","extract_error":null,"extract_retries":0,"content_length":1170,"summary_length":1170}},"news_item":{"id":94896,"canonical_url":"https://tech.huanqiu.com/article/4TSPbwKRYkK","source_url":"https://tech.huanqiu.com/article/4TSPbwKRYkK","title":"对标GPT与Claude，谷歌新大模型暂不面向普通用户 - 环球网","source_name":"环球网","author":null,"published_at":"2026-10-02T10:00:00+00:00","locale":"zh","topic":"ai","tags":[],"rss_summary":"<a href=\"https://news.google.com/rss/articles/CBMiV0FVX3lxTFBVMVpZLWo1b3JmcVlWSjdyY3RvZ3k1OFdqNW4tU2Qza3dIRmkxX212Yy12R1FJLUEtbEFiNjB5R2xHWnllblZTT1lWLVYxWkJYM043RDhYTQ?oc=5\" target=\"_blank\">对标GPT与Claude，谷歌新大模型暂不面向普通用户</a>&nbsp;&nbsp;<font color=\"#6f6f6f\">环球网</font>","full_text":"4TSPbwKRYkK tech.huanqiu.comarticle对标GPT与Claude，谷歌新大模型暂不面向普通用户/e3pmh164r/e3pmtmdvg<article><section data-type=\"rtext\"><p>【环球网科技综合报道】10月3日，据外媒CNBC消息，谷歌正式发布新一代前沿旗舰大模型Gemini 4 Argon，直面OpenAI、Anthropic两大对手的竞争。外媒指出，谷歌选择采取受控分阶段的上线策略，暂不对普通用户开放，优先面向网络安全领域经过筛选的专业人员。</p><p><i class=\"pic-con\"><img data-alt=\"\" src=\"//img.huanqiucdn.cn/dp/api/files/imageDir/f1c9d4e4b736f5a9f5eb2309285492ffu1.png?imageView2/2/w/1260\" /></i></p><p>据介绍，Gemini 4 Argon重点强化长周期任务推理、软件工程、企业知识处理与网络安全防御能力，输出Token上限由上代6.4万提升至100万，能够一次性处理数十万Token的复杂任务，已经在谷歌内部用于大型代码库迁移、科研研发等实际工作。在官方19项基准测试中，该模型拿下13项领先成绩。</p> <adv-loader __attr__inner=\"7004636\" __attr__style=\"width: auto;position: relative;float: left;border: 1px solid #ebebeb; padding: 20px;overflow: hidden;margin: 10px 30px 40px 0;\"></adv-loader> <p>在上线节奏上，该模型通过Fairwind项目，仅向经过严格审核的网络安全从业者开放，同时参与美国政府的预发布安全评估。推广期API定价为每百万输入Token2美元、输出 Token10美元，缓存输入可享受大幅折扣，后续将上调价格。面向普通开发者与消费者的全面开放时间暂未公布。</p><p>外媒报道称，近期OpenAI、Anthropic接连曝出前沿大模型安全隐患，AI安全成为行业焦点。分析师表示，谷歌采取保守受控发布模式，意在塑造安全可信的AI厂商定位。但业内也保持理性，基准跑分亮眼不等于实际业务场景同样强大。（旺旺）</p></section></article>1790989252039环球网版权作品，未经书面授权，严禁转载或镜像，违者将被追究法律责任。责编：王楠环球网179098925203911[]{\"email\":\"wangnan@huanqiu.com\",\"name\":\"王楠\"}","excerpt":"4TSPbwKRYkK tech.huanqiu.comarticle对标GPT与Claude，谷歌新大模型暂不面向普通用户/e3pmh164r/e3pmtmdvg<article><section data-type=\"rtext\"><p>【环球网科技综合报道】10月3日，据外媒CNBC消息，谷歌正式发布新一代前沿旗舰大模型Gemini 4 Argon，直面OpenAI、Anthropic两大对手的竞争。外媒指出，谷歌选择采取受控分阶段的上线策略，暂不对普通用户开放，优先面向网络安全领域经过筛选的专业人员。</p><p><i class=\"pic-con\"><img data-alt=\"\" src=\"//img.huanqiucdn.cn/dp/api/files/imageDir/f1c9d4e4b736f5a9f5eb2309285492ffu1.png?imageView2/2/w/1260\" /></i></p><p>据介绍，Gemini 4 Argon重点强化长周期任务推理、软件工程、企业知识处理与网络安全防御能力，输出Token上限由上代6.4万提升至100万，能够一次性处理数十万Token的复杂任务，已经在谷歌内部用于大型代码库迁移、科研研发等实际工作。在官方19项基准测试中，该模型拿下13项领先成绩。</p> <adv-loader __attr__inner=\"7004636\" __attr__style=\"width: auto;position: relative;float: left;border: 1px solid #ebebeb; padding: 20px;overflow: hidden;margin: 10px 30px 40px 0;\"></adv-loader> <p>在上线节奏上，该模型通过Fairwind项目，仅向经过严格审核的网络安全从业者开放，同时参与美国政府的预发布安全评估。推广期API定价为每百万输入Token2美元、输出 Token10美元，缓存输入可享受大幅折扣，后续将上调价格。面向普通开发者与消费者的全面开放时间暂未公布。</p><p>外媒报道称，近期OpenAI、Anthropic接连曝出前沿大模型安全隐患，AI安全成为行业焦点。分析师表示，谷歌采取保守受控发布模式，意在塑造安全可信的AI厂商定位。但业内也保持理性，基准跑分亮眼不等于实际业务场景同样强大。（旺旺）</p></section></article>1790989252039环球网版权作品，未经书面授权，严禁转载或镜像，违者将被追究法律责任。责编：王楠环球网179098925203911[]{\"email\":\"wangnan@huanqiu.com\",\"name\":\"王楠\"}","extraction":{"state":"ok","confidence":0.9,"error":null,"explanation":"High confidence: full text extraction produced 1170 characters.","diagnostics_url":"/api/diagnose?url=https%3A//tech.huanqiu.com/article/4TSPbwKRYkK","quality_profile":{"profile_version":"extraction_quality.v2","bucket":"high","confidence":0.9,"failure_kind":"none","retryable":false,"retry_after_attempts":0,"reason":"High confidence: full text extraction produced 1170 characters.","operator_guidance":{"severity":"ok","recommended_action":"trust_full_text","next_step":"Use the extracted full text as the primary article source.","operator_label":"Ready","can_retry":false,"can_use_summary":false,"diagnostics_required":false},"content_depth":{"contract_version":"content_depth.v1","category":"full_text","label":"Full text","has_full_text":true,"has_summary":true,"content_length":1170,"summary_length":1170,"usable_text_length":1170,"source_field":"content"},"legacy_collapsed":false,"signals":{"extract_state":"ok","extract_error":null,"extract_retries":0,"content_length":1170,"summary_length":1170}}},"display_formats":["compact","card","full","digest_section","json"]},"daily_stack_record":{"title":"对标GPT与Claude，谷歌新大模型暂不面向普通用户 - 环球网","url":"https://tech.huanqiu.com/article/4TSPbwKRYkK","summary":"4TSPbwKRYkK tech.huanqiu.comarticle对标GPT与Claude，谷歌新大模型暂不面向普通用户/e3pmh164r/e3pmtmdvg<article><section data-type=\"rtext\"><p>【环球网科技综合报道】10月3日，据外媒CNBC消息，谷歌正式发布新一代前沿旗舰大模型Gemini 4 Argon，直面OpenAI、Anthropic两大对手的竞争。外媒指出，谷歌选择采取受控分阶段的上线策略，暂不对普通用户开放，优先面向网络安全领域经过筛选的专业人员。</p><p><i class=\"pic-con\"><img data-alt=\"\" src=\"//img.huanqiucdn.cn/dp/api/files/imageDir/f1c9d4e4b736f5a9f5eb2309285492ffu1.png?imageView2/2/w/1260\" /></i></p><p>据介绍，Gemini 4 Argon重点强化长周期任务推理、软件工程、企业知识处理与网络安全防御能力，输出Token上限由上代6.4万提升至100万，能够一次性处理数十万Token的复杂任务，已经在谷歌内部用于大型代码库迁移、科研研发等实际工作。在官方19项基准测试中，该模型拿下13项领先成绩。</p> <adv-loader __attr__inner=\"7004636\" __attr__style=\"width: auto;position: relative;float: left;border: 1px solid #ebebeb; padding: 20px;overflow: hidden;margin: 10px 30px 40px 0;\"></adv-loader> <p>在上线节奏上，该模型通过Fairwind项目，仅向经过严格审核的网络安全从业者开放，同时参与美国政府的预发布安全评估。推广期API定价为每百万输入Token2美元、输出 Token10美元，缓存输入可享受大幅折扣，后续将上调价格。面向普通开发者与消费者的全面开放时间暂未公布。</p><p>外媒报道称，近期OpenAI、Anthropic接连曝出前沿大模型安全隐患，AI安全成为行业焦点。分析师表示，谷歌采取保守受控发布模式，意在塑造安全可信的AI厂商定位。但业内也保持理性，基准跑分亮眼不等于实际业务场景同样强大。（旺旺）</p></section></article>1790989252039环球网版权作品，未经书面授权，严禁转载或镜像，违者将被追究法律责任。责编：王楠环球网179098925203911[]{\"email\":\"wangnan@huanqiu.com\",\"name\":\"王楠\"}","source":"环球网","date":"2026-10-02T10:00:00+00:00","content":"4TSPbwKRYkK tech.huanqiu.comarticle对标GPT与Claude，谷歌新大模型暂不面向普通用户/e3pmh164r/e3pmtmdvg<article><section data-type=\"rtext\"><p>【环球网科技综合报道】10月3日，据外媒CNBC消息，谷歌正式发布新一代前沿旗舰大模型Gemini 4 Argon，直面OpenAI、Anthropic两大对手的竞争。外媒指出，谷歌选择采取受控分阶段的上线策略，暂不对普通用户开放，优先面向网络安全领域经过筛选的专业人员。</p><p><i class=\"pic-con\"><img data-alt=\"\" src=\"//img.huanqiucdn.cn/dp/api/files/imageDir/f1c9d4e4b736f5a9f5eb2309285492ffu1.png?imageView2/2/w/1260\" /></i></p><p>据介绍，Gemini 4 Argon重点强化长周期任务推理、软件工程、企业知识处理与网络安全防御能力，输出Token上限由上代6.4万提升至100万，能够一次性处理数十万Token的复杂任务，已经在谷歌内部用于大型代码库迁移、科研研发等实际工作。在官方19项基准测试中，该模型拿下13项领先成绩。</p> <adv-loader __attr__inner=\"7004636\" __attr__style=\"width: auto;position: relative;float: left;border: 1px solid #ebebeb; padding: 20px;overflow: hidden;margin: 10px 30px 40px 0;\"></adv-loader> <p>在上线节奏上，该模型通过Fairwind项目，仅向经过严格审核的网络安全从业者开放，同时参与美国政府的预发布安全评估。推广期API定价为每百万输入Token2美元、输出 Token10美元，缓存输入可享受大幅折扣，后续将上调价格。面向普通开发者与消费者的全面开放时间暂未公布。</p><p>外媒报道称，近期OpenAI、Anthropic接连曝出前沿大模型安全隐患，AI安全成为行业焦点。分析师表示，谷歌采取保守受控发布模式，意在塑造安全可信的AI厂商定位。但业内也保持理性，基准跑分亮眼不等于实际业务场景同样强大。（旺旺）</p></section></article>1790989252039环球网版权作品，未经书面授权，严禁转载或镜像，违者将被追究法律责任。责编：王楠环球网179098925203911[]{\"email\":\"wangnan@huanqiu.com\",\"name\":\"王楠\"}","confidence":0.9,"diagnostics_url":"/api/diagnose?url=https%3A//tech.huanqiu.com/article/4TSPbwKRYkK","quality_bucket":"high","failure_kind":"none","retryable":false,"quality_reason":"High confidence: full text extraction produced 1170 characters.","quality_profile":{"profile_version":"extraction_quality.v2","bucket":"high","confidence":0.9,"failure_kind":"none","retryable":false,"retry_after_attempts":0,"reason":"High confidence: full text extraction produced 1170 characters.","operator_guidance":{"severity":"ok","recommended_action":"trust_full_text","next_step":"Use the extracted full text as the primary article source.","operator_label":"Ready","can_retry":false,"can_use_summary":false,"diagnostics_required":false},"content_depth":{"contract_version":"content_depth.v1","category":"full_text","label":"Full text","has_full_text":true,"has_summary":true,"content_length":1170,"summary_length":1170,"usable_text_length":1170,"source_field":"content"},"legacy_collapsed":false,"signals":{"extract_state":"ok","extract_error":null,"extract_retries":0,"content_length":1170,"summary_length":1170}},"tags":[]},"fallback_formats":["markdown","json","html"],"actions":{"read":"/item/94896","export_markdown":"/api/items/94896/export?format=markdown","export_json":"/api/items/94896/export?format=json","diagnose":"/api/diagnose?url=https%3A//tech.huanqiu.com/article/4TSPbwKRYkK"},"formats":{"full":{"id":94896,"title":"对标GPT与Claude，谷歌新大模型暂不面向普通用户 - 环球网","url":"https://tech.huanqiu.com/article/4TSPbwKRYkK","source":"环球网","author":null,"published_at":"2026-10-02T10:00:00+00:00","locale":"zh","topic":"ai","tags":[],"excerpt":"4TSPbwKRYkK tech.huanqiu.comarticle对标GPT与Claude，谷歌新大模型暂不面向普通用户/e3pmh164r/e3pmtmdvg<article><section data-type=\"rtext\"><p>【环球网科技综合报道】10月3日，据外媒CNBC消息，谷歌正式发布新一代前沿旗舰大模型Gemini 4 Argon，直面OpenAI、Anthropic两大对手的竞争。外媒指出，谷歌选择采取受控分阶段的上线策略，暂不对普通用户开放，优先面向网络安全领域经过筛选的专业人员。</p><p><i class=\"pic-con\"><img data-alt=\"\" src=\"//img.huanqiucdn.cn/dp/api/files/imageDir/f1c9d4e4b736f5a9f5eb2309285492ffu1.png?imageView2/2/w/1260\" /></i></p><p>据介绍，Gemini 4 Argon重点强化长周期任务推理、软件工程、企业知识处理与网络安全防御能力，输出Token上限由上代6.4万提升至100万，能够一次性处理数十万Token的复杂任务，已经在谷歌内部用于大型代码库迁移、科研研发等实际工作。在官方19项基准测试中，该模型拿下13项领先成绩。</p> <adv-loader __attr__inner=\"7004636\" __attr__style=\"width: auto;position: relative;float: left;border: 1px solid #ebebeb; padding: 20px;overflow: hidden;margin: 10px 30px 40px 0;\"></adv-loader> <p>在上线节奏上，该模型通过Fairwind项目，仅向经过严格审核的网络安全从业者开放，同时参与美国政府的预发布安全评估。推广期API定价为每百万输入Token2美元、输出 Token10美元，缓存输入可享受大幅折扣，后续将上调价格。面向普通开发者与消费者的全面开放时间暂未公布。</p><p>外媒报道称，近期OpenAI、Anthropic接连曝出前沿大模型安全隐患，AI安全成为行业焦点。分析师表示，谷歌采取保守受控发布模式，意在塑造安全可信的AI厂商定位。但业内也保持理性，基准跑分亮眼不等于实际业务场景同样强大。（旺旺）</p></section></article>1790989252039环球网版权作品，未经书面授权，严禁转载或镜像，违者将被追究法律责任。责编：王楠环球网179098925203911[]{\"email\":\"wangnan@huanqiu.com\",\"name\":\"王楠\"}","full_text":"4TSPbwKRYkK tech.huanqiu.comarticle对标GPT与Claude，谷歌新大模型暂不面向普通用户/e3pmh164r/e3pmtmdvg<article><section data-type=\"rtext\"><p>【环球网科技综合报道】10月3日，据外媒CNBC消息，谷歌正式发布新一代前沿旗舰大模型Gemini 4 Argon，直面OpenAI、Anthropic两大对手的竞争。外媒指出，谷歌选择采取受控分阶段的上线策略，暂不对普通用户开放，优先面向网络安全领域经过筛选的专业人员。</p><p><i class=\"pic-con\"><img data-alt=\"\" src=\"//img.huanqiucdn.cn/dp/api/files/imageDir/f1c9d4e4b736f5a9f5eb2309285492ffu1.png?imageView2/2/w/1260\" /></i></p><p>据介绍，Gemini 4 Argon重点强化长周期任务推理、软件工程、企业知识处理与网络安全防御能力，输出Token上限由上代6.4万提升至100万，能够一次性处理数十万Token的复杂任务，已经在谷歌内部用于大型代码库迁移、科研研发等实际工作。在官方19项基准测试中，该模型拿下13项领先成绩。</p> <adv-loader __attr__inner=\"7004636\" __attr__style=\"width: auto;position: relative;float: left;border: 1px solid #ebebeb; padding: 20px;overflow: hidden;margin: 10px 30px 40px 0;\"></adv-loader> <p>在上线节奏上，该模型通过Fairwind项目，仅向经过严格审核的网络安全从业者开放，同时参与美国政府的预发布安全评估。推广期API定价为每百万输入Token2美元、输出 Token10美元，缓存输入可享受大幅折扣，后续将上调价格。面向普通开发者与消费者的全面开放时间暂未公布。</p><p>外媒报道称，近期OpenAI、Anthropic接连曝出前沿大模型安全隐患，AI安全成为行业焦点。分析师表示，谷歌采取保守受控发布模式，意在塑造安全可信的AI厂商定位。但业内也保持理性，基准跑分亮眼不等于实际业务场景同样强大。（旺旺）</p></section></article>1790989252039环球网版权作品，未经书面授权，严禁转载或镜像，违者将被追究法律责任。责编：王楠环球网179098925203911[]{\"email\":\"wangnan@huanqiu.com\",\"name\":\"王楠\"}","reading_time_min":1,"extraction":{"state":"ok","confidence":0.9,"error":null,"explanation":"High confidence: full text extraction produced 1170 characters.","diagnostics_url":"/api/diagnose?url=https%3A//tech.huanqiu.com/article/4TSPbwKRYkK","quality_profile":{"profile_version":"extraction_quality.v2","bucket":"high","confidence":0.9,"failure_kind":"none","retryable":false,"retry_after_attempts":0,"reason":"High confidence: full text extraction produced 1170 characters.","operator_guidance":{"severity":"ok","recommended_action":"trust_full_text","next_step":"Use the extracted full text as the primary article source.","operator_label":"Ready","can_retry":false,"can_use_summary":false,"diagnostics_required":false},"content_depth":{"contract_version":"content_depth.v1","category":"full_text","label":"Full text","has_full_text":true,"has_summary":true,"content_length":1170,"summary_length":1170,"usable_text_length":1170,"source_field":"content"},"legacy_collapsed":false,"signals":{"extract_state":"ok","extract_error":null,"extract_retries":0,"content_length":1170,"summary_length":1170}}},"quality_profile":{"profile_version":"extraction_quality.v2","bucket":"high","confidence":0.9,"failure_kind":"none","retryable":false,"retry_after_attempts":0,"reason":"High confidence: full text extraction produced 1170 characters.","operator_guidance":{"severity":"ok","recommended_action":"trust_full_text","next_step":"Use the extracted full text as the primary article source.","operator_label":"Ready","can_retry":false,"can_use_summary":false,"diagnostics_required":false},"content_depth":{"contract_version":"content_depth.v1","category":"full_text","label":"Full text","has_full_text":true,"has_summary":true,"content_length":1170,"summary_length":1170,"usable_text_length":1170,"source_field":"content"},"legacy_collapsed":false,"signals":{"extract_state":"ok","extract_error":null,"extract_retries":0,"content_length":1170,"summary_length":1170}},"actions":{"read":"/item/94896","export_markdown":"/api/items/94896/export?format=markdown","export_json":"/api/items/94896/export?format=json","diagnose":"/api/diagnose?url=https%3A//tech.huanqiu.com/article/4TSPbwKRYkK"}},"digest":{"id":94896,"title":"对标GPT与Claude，谷歌新大模型暂不面向普通用户 - 环球网","url":"https://tech.huanqiu.com/article/4TSPbwKRYkK","source":"环球网","topic":"ai","published_at":"2026-10-02T10:00:00+00:00","excerpt":"4TSPbwKRYkK tech.huanqiu.comarticle对标GPT与Claude，谷歌新大模型暂不面向普通用户/e3pmh164r/e3pmtmdvg<article><section data-type=\"rtext\"><p>【环球网科技综合报道】10月3日，据外媒CNBC消息，谷歌正式发布新一代前沿旗舰大模型Gemini 4 Argon，直面OpenAI、Anthropic两大对手的竞争。外媒指出，谷歌选择采取受控分阶段的上线策略，暂不对普通用户开放，优先面向网络安全领域经过筛选的专业人员。</p><p><i…","quality_bucket":"high","quality_reason":"High confidence: full text extraction produced 1170 characters.","reading_time_min":1,"cluster_id":null},"card":{"display_title":"对标GPT与Claude，谷歌新大模型暂不面向普通用户 - 环球网","subtitle":"环球网 · 2026-10-02","summary":"4TSPbwKRYkK tech.huanqiu.comarticle对标GPT与Claude，谷歌新大模型暂不面向普通用户/e3pmh164r/e3pmtmdvg<article><section data-type=\"rtext\"><p>【环球网科技综合报道】10月3日，据外媒CNBC消息，谷歌正式发布新一代前沿旗舰大模型Gemini 4…","badges":["quality:high"],"links":{"read":"/item/94896","original":"https://tech.huanqiu.com/article/4TSPbwKRYkK","diagnose":"/api/diagnose?url=https%3A//tech.huanqiu.com/article/4TSPbwKRYkK"},"quality_warning":null},"export":{"title":"对标GPT与Claude，谷歌新大模型暂不面向普通用户 - 环球网","url":"https://tech.huanqiu.com/article/4TSPbwKRYkK","summary":"4TSPbwKRYkK tech.huanqiu.comarticle对标GPT与Claude，谷歌新大模型暂不面向普通用户/e3pmh164r/e3pmtmdvg<article><section data-type=\"rtext\"><p>【环球网科技综合报道】10月3日，据外媒CNBC消息，谷歌正式发布新一代前沿旗舰大模型Gemini 4 Argon，直面OpenAI、Anthropic两大对手的竞争。外媒指出，谷歌选择采取受控分阶段的上线策略，暂不对普通用户开放，优先面向网络安全领域经过筛选的专业人员。</p><p><i class=\"pic-con\"><img data-alt=\"\" src=\"//img.huanqiucdn.cn/dp/api/files/imageDir/f1c9d4e4b736f5a9f5eb2309285492ffu1.png?imageView2/2/w/1260\" /></i></p><p>据介绍，Gemini 4 Argon重点强化长周期任务推理、软件工程、企业知识处理与网络安全防御能力，输出Token上限由上代6.4万提升至100万，能够一次性处理数十万Token的复杂任务，已经在谷歌内部用于大型代码库迁移、科研研发等实际工作。在官方19项基准测试中，该模型拿下13项领先成绩。</p> <adv-loader __attr__inner=\"7004636\" __attr__style=\"width: auto;position: relative;float: left;border: 1px solid #ebebeb; padding: 20px;overflow: hidden;margin: 10px 30px 40px 0;\"></adv-loader> <p>在上线节奏上，该模型通过Fairwind项目，仅向经过严格审核的网络安全从业者开放，同时参与美国政府的预发布安全评估。推广期API定价为每百万输入Token2美元、输出 Token10美元，缓存输入可享受大幅折扣，后续将上调价格。面向普通开发者与消费者的全面开放时间暂未公布。</p><p>外媒报道称，近期OpenAI、Anthropic接连曝出前沿大模型安全隐患，AI安全成为行业焦点。分析师表示，谷歌采取保守受控发布模式，意在塑造安全可信的AI厂商定位。但业内也保持理性，基准跑分亮眼不等于实际业务场景同样强大。（旺旺）</p></section></article>1790989252039环球网版权作品，未经书面授权，严禁转载或镜像，违者将被追究法律责任。责编：王楠环球网179098925203911[]{\"email\":\"wangnan@huanqiu.com\",\"name\":\"王楠\"}","source":"环球网","date":"2026-10-02T10:00:00+00:00","content":"4TSPbwKRYkK tech.huanqiu.comarticle对标GPT与Claude，谷歌新大模型暂不面向普通用户/e3pmh164r/e3pmtmdvg<article><section data-type=\"rtext\"><p>【环球网科技综合报道】10月3日，据外媒CNBC消息，谷歌正式发布新一代前沿旗舰大模型Gemini 4 Argon，直面OpenAI、Anthropic两大对手的竞争。外媒指出，谷歌选择采取受控分阶段的上线策略，暂不对普通用户开放，优先面向网络安全领域经过筛选的专业人员。</p><p><i class=\"pic-con\"><img data-alt=\"\" src=\"//img.huanqiucdn.cn/dp/api/files/imageDir/f1c9d4e4b736f5a9f5eb2309285492ffu1.png?imageView2/2/w/1260\" /></i></p><p>据介绍，Gemini 4 Argon重点强化长周期任务推理、软件工程、企业知识处理与网络安全防御能力，输出Token上限由上代6.4万提升至100万，能够一次性处理数十万Token的复杂任务，已经在谷歌内部用于大型代码库迁移、科研研发等实际工作。在官方19项基准测试中，该模型拿下13项领先成绩。</p> <adv-loader __attr__inner=\"7004636\" __attr__style=\"width: auto;position: relative;float: left;border: 1px solid #ebebeb; padding: 20px;overflow: hidden;margin: 10px 30px 40px 0;\"></adv-loader> <p>在上线节奏上，该模型通过Fairwind项目，仅向经过严格审核的网络安全从业者开放，同时参与美国政府的预发布安全评估。推广期API定价为每百万输入Token2美元、输出 Token10美元，缓存输入可享受大幅折扣，后续将上调价格。面向普通开发者与消费者的全面开放时间暂未公布。</p><p>外媒报道称，近期OpenAI、Anthropic接连曝出前沿大模型安全隐患，AI安全成为行业焦点。分析师表示，谷歌采取保守受控发布模式，意在塑造安全可信的AI厂商定位。但业内也保持理性，基准跑分亮眼不等于实际业务场景同样强大。（旺旺）</p></section></article>1790989252039环球网版权作品，未经书面授权，严禁转载或镜像，违者将被追究法律责任。责编：王楠环球网179098925203911[]{\"email\":\"wangnan@huanqiu.com\",\"name\":\"王楠\"}","confidence":0.9,"diagnostics_url":"/api/diagnose?url=https%3A//tech.huanqiu.com/article/4TSPbwKRYkK","quality_bucket":"high","failure_kind":"none","retryable":false,"quality_reason":"High confidence: full text extraction produced 1170 characters.","quality_profile":{"profile_version":"extraction_quality.v2","bucket":"high","confidence":0.9,"failure_kind":"none","retryable":false,"retry_after_attempts":0,"reason":"High confidence: full text extraction produced 1170 characters.","operator_guidance":{"severity":"ok","recommended_action":"trust_full_text","next_step":"Use the extracted full text as the primary article source.","operator_label":"Ready","can_retry":false,"can_use_summary":false,"diagnostics_required":false},"content_depth":{"contract_version":"content_depth.v1","category":"full_text","label":"Full text","has_full_text":true,"has_summary":true,"content_length":1170,"summary_length":1170,"usable_text_length":1170,"source_field":"content"},"legacy_collapsed":false,"signals":{"extract_state":"ok","extract_error":null,"extract_retries":0,"content_length":1170,"summary_length":1170}},"tags":[],"format_contract_version":"news_item_formats.v1"}}}