{"id":79134,"topic":"ai","source":"Sohu","title":"DeepSeek今日降价，但大模型竞争不再只拼API单价 - Sohu","url":"https://m.sohu.com/a/1074180178_250147?scm=10001.325_13-325_13.0.0-0-0-0-0.5_1334","url_hash":"45cd39c84afe2171893e9ba1f1ea84c6dfde10c7","author":"","summary":"<a href=\"https://news.google.com/rss/articles/CBMiiAFBVV95cUxOa0hDLXJuY0VoRzNVMzNTZTNYUS03YzZGZFkwejBydlhkbWtuSkhpaDIyLXg0ckVrNEFKQXgxdW44X2N6Q1pvSm80NHAwaEktWmpOa0MyUHhpVm5Xb2UwYUthVzZHUWVDZEFXZWRlVG1pMjZHWkNBOHJUTWl3X3ZJbmJQaDF6M1k3?oc=5\" target=\"_blank\">DeepSeek今日降价，但大模型竞争不再只拼API单价</a>&nbsp;&nbsp;<font color=\"#6f6f6f\">Sohu</font>","content":"DeepSeek今日降价，但大模型竞争不再只拼API单价\n（图源：视觉中国）\n蓝鲸新闻9月10日讯（记者 卢思叶）9月9日，DeepSeek连续更新多个动态。DeepSeek开放平台先是公告，自9月10日12时起将对Flash系列模型降价，后又宣布10日前后将正式发布V4.1 Flash模型，在性能、费用、速度、总用时等各项指标上全面超越V4 Pro。\n具体看这次降价，空闲时段输入缓存命中价格由此前每百万Tokens 0.05元降至0.02元，降幅60%；缓存未命中由1.5元降至1元，降幅约33%；输出则由4.5元降至4元，降幅约11%。高峰时段价格仍为空闲时段的2倍。\n这次价格调整距离8月13日DeepSeek V4 Pro正式版上线时的涨价，还不到一个月，DeepSeek为何又再次调价？\n蓝鲸科技记者与开发者及企业人士交流发现，模型API的单价已经越来越难代表真实使用成本，用户都偏好价格更低的模型，但模型能力、稳定性不足带来的重复调用、返工和错误修复，同样会迅速推高Token消耗，从而增加成本。","image_url":"//q3.itc.cn/q_70/images03/20260910/b65cebe9bf2b428280c833a43afb4860.jpeg","lang":"zh","published_at":"2026-09-10T03:24:03+00:00","fetched_at":"2026-09-10T04:15:08+00:00","status":"read","starred":0,"extract_state":"ok","summary_auto":"DeepSeek今日降价，但大模型竞争不再只拼API单价\n（图源：视觉中国）\n蓝鲸新闻9月10日讯（记者 卢思叶）9月9日，DeepSeek连续更新多个动态。DeepSeek开放平台先是公告，自9月10日12时起将对Flash系列模型降价，后又宣布10日前后将正式发布V4.1 Flash模型，在性能、费用、速度、总用时等各项指标上全面超越V4 Pro。\n具体看这次降价，空闲时段输入缓存命中价格由此前每百万Tokens 0.05元降至0.02元，降幅60%；缓存未命中由1.5元降至1元，降幅约33%；输出则由4.5元降至4元，降幅约11%。高峰时段价格仍为空闲时段的2倍。\n这次价格调整距离8月13日DeepSeek V4 Pro正式版上线时的涨价，还不到一个月，DeepSeek为何又再次调价？\n蓝鲸科技记者与开发者及企业人士交流发现，模型API的单价已经越来越难代表真实使用成本，用户都偏好价格更低的模型，但模型能力、稳定性不足带来的重复调用、返工和错误修复，同样会迅速推高Token消耗，从而增加成本。","cluster_id":3525232,"extract_retries":0,"extract_error":null,"contract_version":"news_item.v1","format_contract_version":"news_item_formats.v1","dedup_url":"https://m.sohu.com/a/1074180178_250147?scm=10001.325_13-325_13.0.0-0-0-0-0.5_1334","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 456 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":456,"summary_length":456,"usable_text_length":456,"source_field":"content"},"legacy_collapsed":false,"signals":{"extract_state":"ok","extract_error":null,"extract_retries":0,"content_length":456,"summary_length":456}},"news_item":{"id":79134,"canonical_url":"https://m.sohu.com/a/1074180178_250147?scm=10001.325_13-325_13.0.0-0-0-0-0.5_1334","source_url":"https://m.sohu.com/a/1074180178_250147?scm=10001.325_13-325_13.0.0-0-0-0-0.5_1334","title":"DeepSeek今日降价，但大模型竞争不再只拼API单价 - Sohu","source_name":"Sohu","author":null,"published_at":"2026-09-10T03:24:03+00:00","locale":"zh","topic":"ai","tags":[],"rss_summary":"<a href=\"https://news.google.com/rss/articles/CBMiiAFBVV95cUxOa0hDLXJuY0VoRzNVMzNTZTNYUS03YzZGZFkwejBydlhkbWtuSkhpaDIyLXg0ckVrNEFKQXgxdW44X2N6Q1pvSm80NHAwaEktWmpOa0MyUHhpVm5Xb2UwYUthVzZHUWVDZEFXZWRlVG1pMjZHWkNBOHJUTWl3X3ZJbmJQaDF6M1k3?oc=5\" target=\"_blank\">DeepSeek今日降价，但大模型竞争不再只拼API单价</a>&nbsp;&nbsp;<font color=\"#6f6f6f\">Sohu</font>","full_text":"DeepSeek今日降价，但大模型竞争不再只拼API单价\n（图源：视觉中国）\n蓝鲸新闻9月10日讯（记者 卢思叶）9月9日，DeepSeek连续更新多个动态。DeepSeek开放平台先是公告，自9月10日12时起将对Flash系列模型降价，后又宣布10日前后将正式发布V4.1 Flash模型，在性能、费用、速度、总用时等各项指标上全面超越V4 Pro。\n具体看这次降价，空闲时段输入缓存命中价格由此前每百万Tokens 0.05元降至0.02元，降幅60%；缓存未命中由1.5元降至1元，降幅约33%；输出则由4.5元降至4元，降幅约11%。高峰时段价格仍为空闲时段的2倍。\n这次价格调整距离8月13日DeepSeek V4 Pro正式版上线时的涨价，还不到一个月，DeepSeek为何又再次调价？\n蓝鲸科技记者与开发者及企业人士交流发现，模型API的单价已经越来越难代表真实使用成本，用户都偏好价格更低的模型，但模型能力、稳定性不足带来的重复调用、返工和错误修复，同样会迅速推高Token消耗，从而增加成本。","excerpt":"DeepSeek今日降价，但大模型竞争不再只拼API单价\n（图源：视觉中国）\n蓝鲸新闻9月10日讯（记者 卢思叶）9月9日，DeepSeek连续更新多个动态。DeepSeek开放平台先是公告，自9月10日12时起将对Flash系列模型降价，后又宣布10日前后将正式发布V4.1 Flash模型，在性能、费用、速度、总用时等各项指标上全面超越V4 Pro。\n具体看这次降价，空闲时段输入缓存命中价格由此前每百万Tokens 0.05元降至0.02元，降幅60%；缓存未命中由1.5元降至1元，降幅约33%；输出则由4.5元降至4元，降幅约11%。高峰时段价格仍为空闲时段的2倍。\n这次价格调整距离8月13日DeepSeek V4 Pro正式版上线时的涨价，还不到一个月，DeepSeek为何又再次调价？\n蓝鲸科技记者与开发者及企业人士交流发现，模型API的单价已经越来越难代表真实使用成本，用户都偏好价格更低的模型，但模型能力、稳定性不足带来的重复调用、返工和错误修复，同样会迅速推高Token消耗，从而增加成本。","extraction":{"state":"ok","confidence":0.9,"error":null,"explanation":"High confidence: full text extraction produced 456 characters.","diagnostics_url":"/api/diagnose?url=https%3A//m.sohu.com/a/1074180178_250147%3Fscm%3D10001.325_13-325_13.0.0-0-0-0-0.5_1334","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 456 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":456,"summary_length":456,"usable_text_length":456,"source_field":"content"},"legacy_collapsed":false,"signals":{"extract_state":"ok","extract_error":null,"extract_retries":0,"content_length":456,"summary_length":456}}},"display_formats":["compact","card","full","digest_section","json"]},"daily_stack_record":{"title":"DeepSeek今日降价，但大模型竞争不再只拼API单价 - Sohu","url":"https://m.sohu.com/a/1074180178_250147?scm=10001.325_13-325_13.0.0-0-0-0-0.5_1334","summary":"DeepSeek今日降价，但大模型竞争不再只拼API单价\n（图源：视觉中国）\n蓝鲸新闻9月10日讯（记者 卢思叶）9月9日，DeepSeek连续更新多个动态。DeepSeek开放平台先是公告，自9月10日12时起将对Flash系列模型降价，后又宣布10日前后将正式发布V4.1 Flash模型，在性能、费用、速度、总用时等各项指标上全面超越V4 Pro。\n具体看这次降价，空闲时段输入缓存命中价格由此前每百万Tokens 0.05元降至0.02元，降幅60%；缓存未命中由1.5元降至1元，降幅约33%；输出则由4.5元降至4元，降幅约11%。高峰时段价格仍为空闲时段的2倍。\n这次价格调整距离8月13日DeepSeek V4 Pro正式版上线时的涨价，还不到一个月，DeepSeek为何又再次调价？\n蓝鲸科技记者与开发者及企业人士交流发现，模型API的单价已经越来越难代表真实使用成本，用户都偏好价格更低的模型，但模型能力、稳定性不足带来的重复调用、返工和错误修复，同样会迅速推高Token消耗，从而增加成本。","source":"Sohu","date":"2026-09-10T03:24:03+00:00","content":"DeepSeek今日降价，但大模型竞争不再只拼API单价\n（图源：视觉中国）\n蓝鲸新闻9月10日讯（记者 卢思叶）9月9日，DeepSeek连续更新多个动态。DeepSeek开放平台先是公告，自9月10日12时起将对Flash系列模型降价，后又宣布10日前后将正式发布V4.1 Flash模型，在性能、费用、速度、总用时等各项指标上全面超越V4 Pro。\n具体看这次降价，空闲时段输入缓存命中价格由此前每百万Tokens 0.05元降至0.02元，降幅60%；缓存未命中由1.5元降至1元，降幅约33%；输出则由4.5元降至4元，降幅约11%。高峰时段价格仍为空闲时段的2倍。\n这次价格调整距离8月13日DeepSeek V4 Pro正式版上线时的涨价，还不到一个月，DeepSeek为何又再次调价？\n蓝鲸科技记者与开发者及企业人士交流发现，模型API的单价已经越来越难代表真实使用成本，用户都偏好价格更低的模型，但模型能力、稳定性不足带来的重复调用、返工和错误修复，同样会迅速推高Token消耗，从而增加成本。","confidence":0.9,"diagnostics_url":"/api/diagnose?url=https%3A//m.sohu.com/a/1074180178_250147%3Fscm%3D10001.325_13-325_13.0.0-0-0-0-0.5_1334","quality_bucket":"high","failure_kind":"none","retryable":false,"quality_reason":"High confidence: full text extraction produced 456 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 456 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":456,"summary_length":456,"usable_text_length":456,"source_field":"content"},"legacy_collapsed":false,"signals":{"extract_state":"ok","extract_error":null,"extract_retries":0,"content_length":456,"summary_length":456}},"tags":[]},"fallback_formats":["markdown","json","html"],"actions":{"read":"/item/79134","export_markdown":"/api/items/79134/export?format=markdown","export_json":"/api/items/79134/export?format=json","diagnose":"/api/diagnose?url=https%3A//m.sohu.com/a/1074180178_250147%3Fscm%3D10001.325_13-325_13.0.0-0-0-0-0.5_1334"},"formats":{"full":{"id":79134,"title":"DeepSeek今日降价，但大模型竞争不再只拼API单价 - Sohu","url":"https://m.sohu.com/a/1074180178_250147?scm=10001.325_13-325_13.0.0-0-0-0-0.5_1334","source":"Sohu","author":null,"published_at":"2026-09-10T03:24:03+00:00","locale":"zh","topic":"ai","tags":[],"excerpt":"DeepSeek今日降价，但大模型竞争不再只拼API单价\n（图源：视觉中国）\n蓝鲸新闻9月10日讯（记者 卢思叶）9月9日，DeepSeek连续更新多个动态。DeepSeek开放平台先是公告，自9月10日12时起将对Flash系列模型降价，后又宣布10日前后将正式发布V4.1 Flash模型，在性能、费用、速度、总用时等各项指标上全面超越V4 Pro。\n具体看这次降价，空闲时段输入缓存命中价格由此前每百万Tokens 0.05元降至0.02元，降幅60%；缓存未命中由1.5元降至1元，降幅约33%；输出则由4.5元降至4元，降幅约11%。高峰时段价格仍为空闲时段的2倍。\n这次价格调整距离8月13日DeepSeek V4 Pro正式版上线时的涨价，还不到一个月，DeepSeek为何又再次调价？\n蓝鲸科技记者与开发者及企业人士交流发现，模型API的单价已经越来越难代表真实使用成本，用户都偏好价格更低的模型，但模型能力、稳定性不足带来的重复调用、返工和错误修复，同样会迅速推高Token消耗，从而增加成本。","full_text":"DeepSeek今日降价，但大模型竞争不再只拼API单价\n（图源：视觉中国）\n蓝鲸新闻9月10日讯（记者 卢思叶）9月9日，DeepSeek连续更新多个动态。DeepSeek开放平台先是公告，自9月10日12时起将对Flash系列模型降价，后又宣布10日前后将正式发布V4.1 Flash模型，在性能、费用、速度、总用时等各项指标上全面超越V4 Pro。\n具体看这次降价，空闲时段输入缓存命中价格由此前每百万Tokens 0.05元降至0.02元，降幅60%；缓存未命中由1.5元降至1元，降幅约33%；输出则由4.5元降至4元，降幅约11%。高峰时段价格仍为空闲时段的2倍。\n这次价格调整距离8月13日DeepSeek V4 Pro正式版上线时的涨价，还不到一个月，DeepSeek为何又再次调价？\n蓝鲸科技记者与开发者及企业人士交流发现，模型API的单价已经越来越难代表真实使用成本，用户都偏好价格更低的模型，但模型能力、稳定性不足带来的重复调用、返工和错误修复，同样会迅速推高Token消耗，从而增加成本。","reading_time_min":1,"extraction":{"state":"ok","confidence":0.9,"error":null,"explanation":"High confidence: full text extraction produced 456 characters.","diagnostics_url":"/api/diagnose?url=https%3A//m.sohu.com/a/1074180178_250147%3Fscm%3D10001.325_13-325_13.0.0-0-0-0-0.5_1334","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 456 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":456,"summary_length":456,"usable_text_length":456,"source_field":"content"},"legacy_collapsed":false,"signals":{"extract_state":"ok","extract_error":null,"extract_retries":0,"content_length":456,"summary_length":456}}},"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 456 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":456,"summary_length":456,"usable_text_length":456,"source_field":"content"},"legacy_collapsed":false,"signals":{"extract_state":"ok","extract_error":null,"extract_retries":0,"content_length":456,"summary_length":456}},"actions":{"read":"/item/79134","export_markdown":"/api/items/79134/export?format=markdown","export_json":"/api/items/79134/export?format=json","diagnose":"/api/diagnose?url=https%3A//m.sohu.com/a/1074180178_250147%3Fscm%3D10001.325_13-325_13.0.0-0-0-0-0.5_1334"}},"digest":{"id":79134,"title":"DeepSeek今日降价，但大模型竞争不再只拼API单价 - Sohu","url":"https://m.sohu.com/a/1074180178_250147?scm=10001.325_13-325_13.0.0-0-0-0-0.5_1334","source":"Sohu","topic":"ai","published_at":"2026-09-10T03:24:03+00:00","excerpt":"DeepSeek今日降价，但大模型竞争不再只拼API单价 （图源：视觉中国） 蓝鲸新闻9月10日讯（记者 卢思叶）9月9日，DeepSeek连续更新多个动态。DeepSeek开放平台先是公告，自9月10日12时起将对Flash系列模型降价，后又宣布10日前后将正式发布V4.1 Flash模型，在性能、费用、速度、总用时等各项指标上全面超越V4 Pro。 具体看这次降价，空闲时段输入缓存命中价格由此前每百万Tokens…","quality_bucket":"high","quality_reason":"High confidence: full text extraction produced 456 characters.","reading_time_min":1,"cluster_id":3525232},"card":{"display_title":"DeepSeek今日降价，但大模型竞争不再只拼API单价 - Sohu","subtitle":"Sohu · 2026-09-10","summary":"DeepSeek今日降价，但大模型竞争不再只拼API单价 （图源：视觉中国） 蓝鲸新闻9月10日讯（记者 卢思叶）9月9日，DeepSeek连续更新多个动态。DeepSeek开放平台先是公告，自9月10日12时起将对Flash系列模型降价，后又宣布10日前后将正式发布V4.1 Flash模型，在性能、费用、速度、总用时等各项指标上全面超越V4 Pro。 具体看这次降价，空闲时段输入缓存命中价格由此前每百万Tokens…","badges":["quality:high"],"links":{"read":"/item/79134","original":"https://m.sohu.com/a/1074180178_250147?scm=10001.325_13-325_13.0.0-0-0-0-0.5_1334","diagnose":"/api/diagnose?url=https%3A//m.sohu.com/a/1074180178_250147%3Fscm%3D10001.325_13-325_13.0.0-0-0-0-0.5_1334"},"quality_warning":null},"export":{"title":"DeepSeek今日降价，但大模型竞争不再只拼API单价 - Sohu","url":"https://m.sohu.com/a/1074180178_250147?scm=10001.325_13-325_13.0.0-0-0-0-0.5_1334","summary":"DeepSeek今日降价，但大模型竞争不再只拼API单价\n（图源：视觉中国）\n蓝鲸新闻9月10日讯（记者 卢思叶）9月9日，DeepSeek连续更新多个动态。DeepSeek开放平台先是公告，自9月10日12时起将对Flash系列模型降价，后又宣布10日前后将正式发布V4.1 Flash模型，在性能、费用、速度、总用时等各项指标上全面超越V4 Pro。\n具体看这次降价，空闲时段输入缓存命中价格由此前每百万Tokens 0.05元降至0.02元，降幅60%；缓存未命中由1.5元降至1元，降幅约33%；输出则由4.5元降至4元，降幅约11%。高峰时段价格仍为空闲时段的2倍。\n这次价格调整距离8月13日DeepSeek V4 Pro正式版上线时的涨价，还不到一个月，DeepSeek为何又再次调价？\n蓝鲸科技记者与开发者及企业人士交流发现，模型API的单价已经越来越难代表真实使用成本，用户都偏好价格更低的模型，但模型能力、稳定性不足带来的重复调用、返工和错误修复，同样会迅速推高Token消耗，从而增加成本。","source":"Sohu","date":"2026-09-10T03:24:03+00:00","content":"DeepSeek今日降价，但大模型竞争不再只拼API单价\n（图源：视觉中国）\n蓝鲸新闻9月10日讯（记者 卢思叶）9月9日，DeepSeek连续更新多个动态。DeepSeek开放平台先是公告，自9月10日12时起将对Flash系列模型降价，后又宣布10日前后将正式发布V4.1 Flash模型，在性能、费用、速度、总用时等各项指标上全面超越V4 Pro。\n具体看这次降价，空闲时段输入缓存命中价格由此前每百万Tokens 0.05元降至0.02元，降幅60%；缓存未命中由1.5元降至1元，降幅约33%；输出则由4.5元降至4元，降幅约11%。高峰时段价格仍为空闲时段的2倍。\n这次价格调整距离8月13日DeepSeek V4 Pro正式版上线时的涨价，还不到一个月，DeepSeek为何又再次调价？\n蓝鲸科技记者与开发者及企业人士交流发现，模型API的单价已经越来越难代表真实使用成本，用户都偏好价格更低的模型，但模型能力、稳定性不足带来的重复调用、返工和错误修复，同样会迅速推高Token消耗，从而增加成本。","confidence":0.9,"diagnostics_url":"/api/diagnose?url=https%3A//m.sohu.com/a/1074180178_250147%3Fscm%3D10001.325_13-325_13.0.0-0-0-0-0.5_1334","quality_bucket":"high","failure_kind":"none","retryable":false,"quality_reason":"High confidence: full text extraction produced 456 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 456 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":456,"summary_length":456,"usable_text_length":456,"source_field":"content"},"legacy_collapsed":false,"signals":{"extract_state":"ok","extract_error":null,"extract_retries":0,"content_length":456,"summary_length":456}},"tags":[],"format_contract_version":"news_item_formats.v1"}}}