← All topics

DeepSeek news, model and product timeline

79 company and product stories · 61 issues


DeepSeek develops large language models and provides access through its website, app and API. Harness offers an environment for tasks and plugins, while open-source projects let developers study and adapt related tools. Follow model updates, product previews, research prototypes and business developments here.

Editorial review · Coverage through

Recent coverage

Selected milestones

Dates refer to our coverage. Each milestone retains the reported stage at that time. Follow the news timeline for subsequent developments.

  1. V4.1-Flash-1017 update

    DeepSeek updates V4.1-Flash-1017, improving project development and shortening task completion time

    Read original coverage ↗
  2. Super-Agent research prototype released

    DeepSeek open-sources a Super-Agent prototype for team coding workflows

    Read original coverage ↗
  3. Harness Creator Mode

    DeepSeek showcases Harness Creator Mode for building reusable plugins through chat

    Read original coverage ↗

Topic guides

Financing and transaction coverage

Dates are our publication dates. Announcements and talks are not closing; different reports may concern the same transaction.

  1. · DeepSeek · Undisclosed (one fund: RMB 2.9 billion) · First external round · Filing or placement
  2. · DeepSeek · About RMB 106 billion (original report) · Second round (reported closed) · Reported

Coverage & briefs

Filter complete history · 82 entries

IT之家 reports DeepSeek front-end code hints at quota management and a possible “Coding Plan”

IT之家 reported on Nov. 14 that developers found quota-management logic in the chat.deepseek.com front end, including a message saying usage limits have been reached and users should try again after quotas recover. The code also appears to support a 5-minute polling cycle while the page is visible to sync model quotas, mute status and server-side copy. The clues suggest a dedicated “Coding Plan,” but DeepSeek has not announced any pricing, tiers or launch date, and the information remains community speculation.

AI-generated

TechCrunch: Research says some Chinese frontier models show signs of Blackwell GPU training

TechCrunch reported that a new research effort led by former OpenAI researcher Will DePue found fingerprints of Blackwell GPU training in open weights from DeepSeek, Moonshot AI, MiniMax and Ant Group. The strongest signals were seen in Moonshot’s Kimi K3 releases and DeepSeek’s V4 models, though the findings do not prove the models were trained entirely on Nvidia chips or that export laws were broken. The report also said Jensen Huang left Nvidia this week amid a scandal involving Megaspeed and possible export-control liability.

AI-generated

IT之家: DeepSeek opens full-suite app APIs with search, files and voice tools

According to IT之家, DeepSeek said on November 13 that it has officially opened its all-in-one app API. Developers can enable new parameters in the existing API to add deep-thinking search, file input and other chat features, while voice and low-level capabilities such as OCR, ASR and TTS are available through separate interfaces. DeepSeek also open-sourced a lightweight front-end demo built on the new API and disclosed usage prices for the added tools.

AI-generated
Mentions in other coverage · 32
Volcano Engine prioritizes MaaS commercialization after RMB4.8 billion in first-half revenue
As customers buy video and complex-task services alongside text, product mix and inference costs jointly shape platform revenue and gross profit.
Zhang Yiming tops Hurun's China list at RMB670 billion; Liang Wenfeng debuts in seventh
AI-related capital value is appearing in the ownership stakes of model, compute and robotics founders, with wealth rankings providing one observable snapshot.
Chinese models take OpenRouter’s weekly top five as Step 5 Preview usage rises 30%
Platform usage rankings reveal developer choices, while industry-wide demand and commercial revenue require different evidence.
Reflection AI releases weights for 501-billion-parameter Beam, built to finish tasks with fewer tokens
Open models are competing on a new dimension: not just who is smarter, but how many tokens it takes to get the same job done.
Doubao phone assistant ends its notice period as 153 apps opt out of screen automation
The hardest problem for phone agents is not model capability but who owns the entry point: whoever taps the screen for the user is diverting attention that used to belong to the apps.
Mistral releases its trillion-parameter Large 4, which third-party testing calls the strongest model from outside the US and China
Open models from Europe and America are both measuring themselves against Chinese ones, so on the open-weight track there is no longer any dispute about who leads.
Moonshot AI closes its final pre-IPO round at a valuation of about $50 billion and targets a Hong Kong listing in the first quarter of next year
Revenue going from $300 million to $1 billion in half a year, and valuation from $31.5 billion to $50 billion, shows Chinese model companies are starting to pitch capital markets on revenue instead of user counts.
Alibaba's ModelScope and OSChina's MoArk vie to become China's Hugging Face
Open models are free; the fight is over where developers first download them, because that's often where they first pay for compute.
Moonshot's Kimi K3.1 surfaces on its API platform, reportedly with a 1M-token context and a launch next month
Chinese model makers are moving from chasing a benchmark score to selling one model at several price points, making pricing design as important as capability.
Xiaomi's MiMo-V2.6-Pro tops the open-weight leaderboard, tied with Grok 4.7
$3 million for an open model tied with Grok 4.7 — a phone maker is repricing the entry fee to frontier capability.
OpenRouter-based estimates put Chinese models at 67.46 trillion weekly tokens, with important coverage limits
Token share shows where models are being tried and called; commercial share still depends on who pays and at what price.
StepFun releases Step 5 Preview, a 600B-parameter MoE with a 1M context window
600 billion total against 27 billion active — Chinese models chose the inference bill over the capability ceiling.
Chinese models reportedly reached 46% of U.S. OpenRouter weekly usage at one point, as open distribution expands reach
Chinese models' overseas reach is visible in token volume, but usage share and revenue share remain different measures.
Moonshot AI files confidentially with the HKEX while raising at a $50 billion pre-money valuation
A company whose valuation rose sevenfold in six months is queuing to list — pricing power is being handed from private markets to public ones.
Z.ai's first-half revenue jumps 400% to 954M yuan, still short of its own $200M projection
Revenue up fourfold, stock down sixty percent — that gap is the market repricing the story told by China's model companies.
Zhipu Open-Sources GLM-5.3-Flash; All Public Beta Traffic Ran on Domestic Chip Clusters
The watershed this time isn't benchmark scores — it's that a frontier-scale model has, for the first time, proven it can survive without NVIDIA's supply rhythm.
Alibaba Open-Sources Qwen3.8-Flash-Next — 125B Model Activates Only 6B Parameters per Token
The one-ninth training-cost figure says more than any benchmark score about which battle the next-generation Qwen aims to win.
Reuters Exclusive: Moonshot AI Seeks 30% Cut From Big Three Cloud Providers for Kimi K3 Hosting
The revenue-share talks were never just about the money — they're about whether Chinese models get a named, legitimate supplier identity on US clouds.
OpenAI Reveals First Benchmarks for In-House Inference Chip Jalapeño — Per-Watt Performance Beats Nvidia GB300
A first-gen ASIC beating a current flagship GPU isn't a process-node win — it's the ability to put models, compilers, and silicon into the same iteration cycle.
Nvidia Groq 3 LPX Racks Enter Full Production, Nebius First AI Cloud to Adopt
By splitting prefill and decode across two chip types, Nvidia concedes that the era of a single GPU handling the entire inference pipeline is over.
ByteDance Folds Trae and Coze Into Doubao, Launches Doubao Work This Week to Take On Tencent's WorkBuddy
ByteDance is moving its org structure, not product features — a sign it sees office AI as a battle that demands concentrated firepower.
OpenAI cuts GPT-5.6 Sol input price to $4 per million tokens, down over 20%
A three-month limited-time price cut is not a promotion — it's OpenAI leaving itself an opening to adjust prices again at any time.
Tencent's Next-Gen Model Hy4 Appears in Internal Test Interface, Parameter Count to Exceed Hy3's 295B
With self-developed and open-source models sitting side by side in the internal test interface, Tencent has no intention of selecting just one of the two paths.
Unitree Technology Opens 629% Higher on STAR Market Debut, Closes Up 460%; Market Cap Tops 340 Billion Yuan
DJI's 10.1286 million yuan in unpaid capital contribution is the most expensive "think again" in this wave of embodied intelligence.
xAI Releases Grok 4.6, Intelligence Index Ties GPT-5.6 Sol, Price Steady at $2
Intelligence gap of 1 point, price gap of 8x — the high-priced tier needs a new justification.
Fable 5 Captures 11.4% of Anthropic Revenue in First Month, Token Volume Just 6%
Usage at 6%, revenue at 11.4% — premium models sell task value, not call counts.
MiniMax Open-Sources Next-Gen Multimodal Model H3, Tops Hugging Face Trending Chart
Eight domestic chips adapted on the same day says more about what MiniMax has a grip on than a No. 1 spot on the leaderboard.
Meta debuts first coding agent Muse Code, with output priced at $4.25 per million tokens
The coding-model price war is on — the first shot lands on competitors' margins, and the ammunition is data.
Alibaba Releases 2.4-Trillion-Parameter Qwen3.8-Max, Says Max-Level Open Weights Next Week
What counts is no longer any one release, but the once-a-month release calendar.
All Top Five in OpenRouter's Weekly Calls Are Chinese Models; Xiaomi MiMo-V2.5 Takes Top Spot with 10.5 Trillion Tokens
Open source isn't a margin concession — it's moving the distribution channel from someone else's console into your own hands.
// related topics — frequently covered alongside #DeepSeek