DeepSeek

A Hangzhou developer of large language modelslarge language model, n.: A computer program trained on enormous amounts of text to predict the words that come next. At vast scale, that simple skill lets it chat, answer questions, translate, summarize and write code., spun out of the Chinese quantitative trading firm High-Flyer, long its only source of money and computing, that publishes its models as open weightsopen-weight model, n.: An AI model whose trained internal numbers are published for anyone to download and run on their own computers, even if the data and code used to make it stay private., mostly under the MIT license. It is known for training efficiency under US export controls on AI chipsexport controls on AI chips, n.: Government rules that limit which countries and companies may buy the most advanced AI chips and the machines used to make them. The United States uses them mainly to slow China's progress in AI.: DeepSeek-V2 (2024) introduced multi-head latent attentionmulti-head latent attention, n.: An attention design from the Chinese lab DeepSeek that compresses the stored information about earlier words into a much smaller summary, so long conversations need far less memory. to shrink the memory needed during generation; DeepSeek-V3 (December 2024) combined fine-grained mixture of expertsmixture of experts, n.: A way to build very large AI models cheaply: the model contains many specialist sub-networks, and for each word only a few are switched on, picked by a small dispatcher network. layers with FP8 training at a reported final-run cost of about $5.6 million in rented GPU time, excluding earlier research; and DeepSeek-R1DeepSeek-R1, prop. n.: A free, downloadable AI reasoning model released by the Chinese company DeepSeek in January 2025. It matched leading American models at math and coding for a fraction of the reported cost, shaking the industry. (January 2025) showed that reinforcement learningreinforcement learning, n.: A way of teaching a computer program by trial and error: it tries actions, earns points for good results and penalties for bad ones, and gradually learns which choices pay off over time. with group relative policy optimizationgroup relative policy optimization, n.: A reinforcement-learning recipe for language models that has the model answer the same question many times and rewards each answer by how much better it did than the group's average. could produce a reasoning modelreasoning model, n.: An AI model trained to work through a problem step by step in hidden 'thinking' text before it answers. Letting it think longer generally makes it better at math, coding and planning. close to OpenAI's o1. R1 topped app-store charts and helped wipe nearly $600 billion from Nvidia's market value in one day, the DeepSeek momentDeepSeek moment, n.: The shock of January 2025, when a little-known Chinese company released a free AI model nearly as good as the best American ones at a fraction of the reported cost, briefly wiping hundreds of billions off chip stocks.. Later models include V3.2 (2025), with sparse attention, and V4, previewed in April 2026. Several governments bar its apps from official devices over data-security concerns.