AI Products / Assistants
DeepSeek
DeepSeek's official chat product, the web and app entry to its open-weights models.DeepSeek 官方对话产品,开放权重模型的网页与 App 入口
DeepSeek 官方对话产品,开放权重模型的网页与 App 入口DeepSeek's official chat product, the web and app entry to its open-weights models.
Ranking history
| List | Period | # | Value | MoM |
|---|---|---|---|---|
| AICPB 全球网站 · 全球 AI 榜 · 网站 | 2026-06 | 5 | 483.80M | -10.59% |
| AICPB 全球App · 全球 AI 榜 · App | 2026-06 | 6 | 139.08M | -1.50% |
| AICPB 中国网站 · 中国 AI 榜 · 网站 | 2026-06 | 1 | 483.80M | -10.59% |
| AICPB 中国App · 中国 AI 榜 · App | 2026-06 | 4 | 139.08M | -1.50% |
| AICPB ChatBot · AI ChatBot 榜 · 网站 | 2026-06 | 4 | 483.80M | -10.59% |
| a16z Web · 第5版 · Web Top 50 | 2025-08 | 3 | — | — |
| a16z App · 第5版 · Mobile Top 50 | 2025-08 | 8 | — | — |
Each row’s methodology is given by the list name (MAU, traffic and token usage differ), so rows are not comparable. AICPB does not disclose its collection method.
Timeline · 34
- X trendThe Information: DeepSeek and Moonshot relied heavily on GPT-6 Astra outputs to train their latest models: Moonshot's Kimi K3.1 Pro (release…The Information:DeepSeek 和月之暗面训练最新模型时大量使用 GPT-6 Astra 的输出:据报道,月之暗面上周发布的 Kimi K3.1 Pro 和 DeepSeek 10 月 17 日的 V4.1-Flash 1017 都受益于蒸馏 Astra;Ope…
- LaunchOpenDsec: an open-source, self-hosted agent sandbox platform built from DeepSeek's DSec paper: an independent project (no DeepSeek code) who…OpenDsec:按 DeepSeek DSec 论文做的开源自托管 Agent 沙箱平台:独立项目(不含 DeepSeek 代码),控制面由无状态 apiserver、基于 Raft 的 IAM、watcher 和每节点 edge 组成,组件间 gRPC + mTLS;节点支持…
- X trendArtificial Analysis launches Intelligence Index v5, with 60% of it now built on private test sets or held-out answers: new entries are Cogni…Artificial Analysis 发布智能指数 v5,60% 的题目改用私有测试集或不公开答案:新增 Cognition 的 FrontierCode 2.0(以维护者是否愿意合并 agent 的改动来评分)和 Terminal-Bench-Science(70 个 age…
- X trendmeng shao: the open-source "ai-app-playbook" has coding agents build profitable consumer AI apps on a fixed stack: human-facing chapters 00–…meng shao:开源「ai-app-playbook」,让 coding agent 按固定栈做出能赚钱的消费级 AI App:给人看的 00–13 章节加给 Agent 执行的 AGENTS.md(Phase 0–14,每步标 [AGENT]/[HUMAN],开工前过 Ga…
- X trendGuizang: the DeepSeek Harness plugin ecosystem is booming: after installing several popular DeepSeek Harness plugins, he notes DeepSeek is c…歸藏:DeepSeek Harness 插件生态正在蓬勃发展:他装了几个热门的 DeepSeek Harness 插件后感慨,DeepSeek 便宜、能力又够强,很多人(尤其爱折腾的年轻人)现在只用 DeepSeek Harness,像当年刷机、折腾 ROM 的那批人。690 赞…
- NewsUbicloud: vLLM's new constrained decoding takes tool-call schema accuracy from 92% to 100%: serving Kimi K2.6 on vLLM 0.28, only 92% of tool…Ubicloud:vLLM 新增的约束解码把工具调用 schema 准确率从 92% 提到 100%:用 vLLM 0.28 部署 Kimi K2.6 时,只有 92% 的工具调用符合预期 schema——每 25 次就有 2 次失败,Agent 随后会在错误上下文上继续推理;v…
- X trendHassan (Together AI): how I build apps these days: record a 1–5 minute voice prompt with Wispr Flow; add visual inspo (screenshots or Inspo…Hassan(Together AI):我现在怎么做应用:用 Wispr Flow 录 1–5 分钟语音当提示词;补视觉参考(截图或 Inspo MCP);用 Matt Pocock 的 grill-me skill 让 Agent 先追问澄清;对 v1 再用语音发反馈;每个新功…
- X trendvista8: a minimal "jailbreak an old Kindle with Codex + install KOReader" tutorial: buy a Ugreen Type-C to Micro-USB cable, charge the Kindl…向阳乔木:用 Codex 给旧 Kindle 越狱并装 KOReader 的极简教程:京东到家买绿联 Type-C 转 Micro-USB 线给旧 Kindle 充电 → 连上电脑跟 Codex 说「帮我越狱 kindle」(拒了就换 DeepSeek)→ 越狱后让它装 KORe…
- X trendStrata: run a 125B-class model locally on a gaming PC and wire it into Claude Code / Codex: an open-source inference engine (MIT, ~32k GitHu…Strata:让普通游戏电脑本地跑 125B 级大模型,并接入 Claude Code / Codex:开源推理引擎(MIT,GitHub 约 3.2 万星),用 GPU 显存、内存和 SSD 三级调度加 MoE 专家动态缓存,显存加内存合计 32GB 起步就能跑 Qwen3.8…
- X trendOpenAI is winning Anthropic coding customers with steep discounts: The Information reports that at Rippling, Anthropic's share of engineers'…OpenAI 靠大幅折扣从 Anthropic 抢走编程客户:据 The Information,Rippling 工程师写代码用的模型里,Anthropic 占比从 5 月的 57% 掉到 18%,OpenAI 从 16% 升到 52%,CEO Parker Conrad 说就…
- NewsV2EX hot thread: "Are people who insist on using Claude masochists?": nearly 100 replies mostly explain seriously why they keep using it des…V2EX 热帖:「硬要用 Claude 的人是不是有 M 倾向?」:近百条回复基本是认真回答为什么顶着封号也要用:有人说同一个提示词 Codex 要 10 分钟、Opus 5.5 medium 只要 1 分钟;有人用 DeepSeek 写的设计文档反复复核没发现问题,Opus 5…
- X trendBloomberg: DeepSeek's new funding round is about ¥100B (about $15B): Tencent and CATL are putting in the most; the target was raised at leas…彭博:DeepSeek 新一轮融资约 1000 亿元人民币(约 150 亿美元):腾讯和宁德时代出资最多,融资目标在谈判中至少上调两次(500 亿 → 800 亿 → 约 1000 亿元);按《财经》报道,投前估值约 5000 亿元没变,相当于新钱按原价买到更多股份。这是 Dee…
- X trendThe MemOS team (MemTensor) open-sources "memory-native" model MemT-1.0, building memory writes, updates and reads into model computation rat…MemOS 团队(MemTensor)开源「记忆原生」模型 MemT-1.0:把记忆的写入、更新和读取做进模型计算,而不是外挂检索。Nano 版 36B 总参数、3B 激活(基于 Qwen3.6-35B-A3B,Apache 2.0),Flash 版 288B、17B 激活(基于…
- X trendDeepSeek open-sources its in-house inference framework DeepInfer (the one behind V4.1 Flash's online service); the official post has 35k lik…DeepSeek 开源自研推理框架 DeepInfer(就是支撑 V4.1 Flash 线上服务的那套),官方帖 3.5 万赞、1.3 万收藏。配套技术报告发在知乎,公开了 Prefill/Decode 硬件选型、KVCache 目标等细节,还给了单集群数据:吞吐 155 万亿…
- X trendThe Information reports Anthropic will release a new model this month with stronger reasoning and "intuition" than current models, especiall…The Information 爆料:Anthropic 本月要发新模型,推理和「直觉」都比现有模型强,特别会揣摩没说出口的意图(例子:让它修一个 bug,它会找出同一类问题并问你要不要一起修),也更少给套路化答案;名字和归属 Fable 还是 Opus 系列未定。同一篇还提到…
- X trendThe US Department of Homeland Security (DHS) ordered civilian federal agencies to remove DeepSeek and "third-party products and services bui…美国国土安全部(DHS)下令联邦民事机构移除 DeepSeek 及「基于 DeepSeek 实现开发的第三方产品和服务」,并禁止采购。命令 7 月 24 日就已签署,10 月 19 日才公开,10 月 20 日登联邦公报;不含国防和情报系统。Rohan Paul 指出即使权重跑在…
- X trendRun DeepSeek V4.1 Flash locally (Unsloth): Unsloth built a llama.cpp implementation and dynamic GGUFs; the 1-bit version shrinks from 508GB…本地跑 DeepSeek V4.1 Flash(Unsloth):Unsloth 做了 llama.cpp 实现和动态 GGUF,1-bit 版本从 508GB 压到 250GB、保留约 88.2% 准确率,128GB+ 内存的机器能跑;Daniel Han 补充 engram/…
- NewsMiniMax M3.1 Flash launches: online inference nearly 3x faster, price down 70%: the anonymous "Space Bunny" that processed 200T+ tokens earl…MiniMax M3.1 Flash 上线:线上推理快了近 3 倍,价格降了 70%:之前匿名处理了 200T 多 token 的「Space Bunny」就是它的早期版本。官方成绩有 DeepSWE v1.1 73.7%、OSWorld V2 61.7 分、Agents' La…
- X trendNetEase LobsterAI publishes the "China Office Agent User Behavior Report", with conclusions quite different from the impression on X: WeChat…网易 LobsterAI 发布《中国办公 Agent 用户行为报告》,结论和 X 上的印象很不一样:微信是最大的办公 agent 入口(77%,飞书 14.3%),最大使用场景仍是泛编程/IT,73.6% 的普通用户只用一个模型,最常见组合是 DeepSeek Flash 搭 G…
- X trendDeepSeek-V4.1-Flash-1017 update: per 智东西, it (Max tier) is already #1 open-source on the Code Arena WebDev leaderboard and #12 overall (1670…DeepSeek-V4.1-Flash-1017 新进展:智东西转述,它(Max 档)已经是 Code Arena WebDev 榜开源第一,总榜第 12(1670 分),排在 Claude Opus 5 (High) 前面,成本约为 Kimi K3 的十分之一。花叔也发了一份价…
- NewsAI data centers, continued: Nscale's IPO valuation gets cut, neoclouds may shrink to just 7–10: following last night's news of Microsoft pul…AI 数据中心续:Nscale IPO 估值遭砍价,neocloud 可能只剩 7–10 家:接着昨晚那期微软退出 Nscale 项目的消息,据消息人士,投资者对 Nscale 约 300 亿美元的 IPO 估值不买账,有人要求降到 160–190 亿美元;The Informa…
- NewsV2EX: SciForge, a local-first deep-learning research agent framework: it splits research into four layers: 24 methodology Skills (compatible…V2EX:SciForge,本地优先的深度学习科研 Agent 框架:把科研拆成四层:24 个方法论 Skill(兼容 Claude Code 规范)、46 个确定性工具(文献检索精读、SSH 跑实验、配图、写作)、7 条流水线(综述、实验、论文、每日跟踪等)和 8 个无构建的本…
- X trendDeepSeek-V4.1-Flash-1017 follow-up: Tencent's WorkBuddy international version is free until 10/31 (no domestic version yet); one developer c…DeepSeek-V4.1-Flash-1017 后续:腾讯 WorkBuddy 国际版限时免费用到 10/31(国内版暂无);有开发者算账说 Claude 和 OpenAI 的 API 联网搜索是每千次 10 美元另加 token,DeepSeek Search 每千次 0.7…
- X trendGive agents cheap web search (向阳乔木): he wrapped DeepSeek's newly opened Search and Fetch APIs (¥5 per thousand calls) into a skill installed…给 agent 接上便宜的联网搜索(向阳乔木):他把 DeepSeek 新开放的 Search 和 Fetch API(每千次 5 元)包成了 skill,用 npx skills add 一行装上,据说还能读公众号文章,可以替代贵的内置搜索。
- LaunchLlmspeed (Show HN): before buying a GPU or Mac, "feel" a local model's output speed in the browser. Pick hardware (RTX 3090/5090, Strix Halo…Llmspeed(Show HN):买显卡或 Mac 之前,先在浏览器里「体感」一下本地模型的出字速度。选硬件(RTX 3090/5090、Strix Halo、Mac Studio M5 Ultra、DGX Spark 等)和模型(Qwen 3.8 27B、DeepSeek V…
- X trendDeepSeek releases DeepSeek-V4.1-Flash-1017 (live at 16:35; official post 1.8M+ views): officially higher benchmark scores, faster task compl…DeepSeek 发布 DeepSeek-V4.1-Flash-1017(16:35 上线,官方帖 1.8M+ 浏览):官方说跑分更高、任务完成更快、输出更短,网页、App、API 已切换,权重放上了 Hugging Face 和 ModelScope;训练用了"带自我反馈的强化…
- NewsMicrosoft releases the MAI-Base-2 foundation model: MoE architecture, 2.8T total parameters with 80B active per token, fully in-house pretra…微软发布 MAI-Base-2 基座模型:MoE 架构,总参数 2.8T、每 token 激活 80B,完全自研预训练。微软称它在内部留出评测上超过所有测过的开放权重模型,预训练 FLOPs 效率是最强开源基座 DeepSeek-V4-Pro-Base 的 4–12 倍,省下的算…
- NewsDeepSeek publishes web search / fetch API pricing: Search and Fetch become separate APIs billed per successful request, each ¥5 per 1000; se…DeepSeek 公开网页搜索 / 抓取 API 定价:Search 和 Fetch 变成独立 API、按成功请求计费,各是每 1000 次 5 元;通过 Claude Code 的 Web Search 或 Codex 的 websearch 发起的搜索,上线后头两周免收 Se…
- X trendThe Information/FT: Anthropic told IPO investors that distillation-type revenue is 5–7% (over $840M in Q3), which it proactively cut in Q3;…The Information/FT:Anthropic 向 IPO 投资者披露蒸馏类收入占 5–7%(Q3 超 8.4 亿美元),Q3 已主动压缩;增速放缓下部分投资者认为估值 1.5 万亿比 2 万亿更合理。
- X trendDeepSeek outage notice: the API was unavailable 10/15 8:11–13:04; compensation is a full credit refund of the seven days of consumption from…DeepSeek 故障说明:API 10/15 8:11–13:04 不可用;补偿为 10/8–10/14 七天消耗全额返赠金。 /
- X trendDeepSeek has been down since 8 am on 10/15 (web/app and the V4.1 Flash/V4 Pro APIs), with third-party relays affected: workbuddy entirely do…DeepSeek 10/15 早 8 点起宕机(网页/App + V4.1 Flash/V4 Pro API),第三方中转连带:workbuddy 全挂、commandcode fast 模型不可用、cline 正常。 /
- NewsDeepSeek web and API partly down this morning: when ITHome reported at 9:06 the official status page already showed an outage of 50+ minutes…DeepSeek 网页和 API 今早部分中断:IT之家 9:06 报道时官方状态页已显示中断 50 多分钟,LinuxDo 最新帖里一大片"DeepSeek 崩了吗"。依赖它 API 的服务可以先看看状态页。IT之家 | LinuxDo
- X trend@guansi: uses Opus 5.5 to direct Codex and hands Codex, DeepSeek, local models and various harnesses to Opus for unified scheduling, saving…@guansi:用 Opus 5.5 指挥 Codex,并把 Codex、DeepSeek、本地模型与各类 Harness 交给 Opus 统一调度,省额度、少动手;适合「强规划模型 + 便宜执行模型」分层。
- X trendLocal inference speed race between GLM 5.3 Flash and DeepSeek V4.1 FlashGLM 5.3 Flash / DeepSeek V4.1 Flash 本地推理速度竞赛