DeepSeek

DeepSeek V3.2模型发布：2026年AI智能体与推理能力全面升级

DeepSeek has released its V3.2 flagship model with enhanced agent capabilities and reasoning integration, available across web, mobile app, and API platforms. 原文翻译：深度求索发布了V3.2旗舰模型，强化了Agent能力和思考推理功能，已在网页端、移动应用和API平台全面上线。

2026/3/2

阅读全文 →

DeepSeek是否从GPT蒸馏而来？2026知识蒸馏技术分析

Knowledge distillation is a model training technique where a smaller student model learns from a larger teacher model, improving efficiency while maintaining performance. This article analyzes whether DeepSeek models were distilled from GPT, examining data, logits, and feature distillation methods. (知识蒸馏是一种模型训练技术，通过教师-学生架构让小模型从大模型中学习知识，在提升效率的同时保持性能。本文深入分析DeepSeek是否从GPT蒸馏而来，探讨数据蒸馏、Logits蒸馏和特征蒸馏三种方法。)

2026/2/16

阅读全文 →

FlashMLA：DeepSeek为Hopper GPU打造的高性能注意力解码内核

FlashMLA is an optimized MLA decoding kernel for Hopper GPUs that significantly improves LLM inference efficiency through advanced attention mechanisms and memory optimization. (FlashMLA是专为Hopper GPU优化的MLA解码内核，通过先进的注意力机制和内存优化显著提升大语言模型推理效率。)

2026/1/24

阅读全文 →

DeepSeek V4前瞻：代码提交揭示下一代AI模型的架构革新与编程能力飞跃

DeepSeek is reportedly developing a new flagship AI model, DeepSeek V4, with enhanced coding capabilities, set to launch around Chinese New Year in mid-February. Recent GitHub code updates reveal a new model identifier "MODEL1" with distinct technical features including KV cache layout, sparsity handling, and FP8 decoding support, suggesting optimized memory and computational efficiency. The model may also incorporate recent research on optimized residual connections and biologically-inspired AI memory modules. (DeepSeek据称正在开发新一代旗舰AI模型DeepSeek V4，具备更强的编程能力，计划于2月中旬农历新年期间发布。近期GitHub代码更新显示新的模型标识符“MODEL1”具有独特技术特征，包括键值缓存布局、稀疏性处理和FP8解码支持，表明在内存优化和计算效率方面进行了针对性设计。该模型可能整合优化残差连接和受生物学启发的AI记忆模块等最新研究成果。)

2026/1/24

阅读全文 →

FlashMLA：DeepSeek为Hopper GPU打造的高性能解码内核

暂无摘要...

2026/1/23

阅读全文 →

DeepSeek发布FlashMLA：专为Hopper GPU优化的高效MLA解码内核，AI推理性能大幅提升

FlashMLA is an efficient MLA decoding kernel optimized for NVIDIA Hopper GPUs, delivering up to 3000 GB/s memory bandwidth and 580 TFLOPS compute performance while reducing KV cache requirements by 93.3% for faster, more cost-effective AI inference. (FlashMLA是DeepSeek针对NVIDIA Hopper GPU优化的高效MLA解码内核，在内存受限配置下可达3000 GB/s带宽，计算受限配置下可达580 TFLOPS峰值性能，同时将KV缓存需求减少93.3%，实现更快、更经济的AI推理。)

2026/1/23

阅读全文 →

DeepSeek FlashMLA代码分析：揭秘未公开的MODEL1高效推理架构

DeepSeek's FlashMLA repository reveals two distinct model architectures: V3.2 optimized for maximum performance and precision, and MODEL1 designed for efficiency and deployability with lower memory footprint and specialized long-sequence handling. (DeepSeek的FlashMLA代码库揭示了两种不同的模型架构：V3.2针对最大性能和精度优化，而MODEL1则针对效率和可部署性设计，具有更低的内存占用和专门的长序列处理能力。)

2026/1/23

阅读全文 →

FlashMLA：DeepSeek开源的高效MLA解码内核，专为NVIDIA Hopper GPU优化

FlashMLA is an open-source, high-performance Multi-Head Linear Attention (MLA) decoding kernel optimized for NVIDIA Hopper architecture GPUs, designed to handle variable-length sequences efficiently. It enhances memory and computational efficiency through optimized KV caching and BF16 data format support, achieving up to 3000 GB/s memory bandwidth and 580 TFLOPS computational performance on H800 SXM5 GPUs. FlashMLA is ideal for large language model (LLM) inference and natural language processing (NLP) tasks requiring efficient decoding. (FlashMLA是DeepSeek开源的高效MLA解码内核，专为NVIDIA Hopper架构GPU优化，用于处理可变长度序列。通过优化KV缓存和采用BF16数据格式，提升了内存和计算效率，在H800 SXM5 GPU上内存带宽可达3000 GB/s，计算性能可达580 TFLOPS。适用于大语言模型推理和需要高效解码的自然语言处理任务。)

2026/1/23

阅读全文 →

DeepSeek开源FlashMLA：面向Hopper GPU的终极解码加速内核，大幅提升大模型推理效率

FlashMLA is an efficient MLA decoding kernel optimized for Hopper GPUs (specifically H800) and variable-length sequences, significantly accelerating inference for large language models. (FlashMLA是一款针对Hopper GPU（特别是H800）和可变长度序列优化的高效MLA解码内核，能大幅加速大语言模型的推理过程。)

2026/1/23

阅读全文 →

FlashMLA：DeepSeek高性能注意力内核库，驱动V3模型实现660 TFLOPS

FlashMLA is DeepSeek's optimized attention kernel library that powers DeepSeek-V3 models, featuring token-level sparse attention with FP8 KV cache support, achieving up to 660 TFLOPS performance on NVIDIA H800 GPUs. (FlashMLA是DeepSeek优化的注意力内核库，为DeepSeek-V3模型提供动力，具有令牌级稀疏注意力和FP8 KV缓存支持，在NVIDIA H800 GPU上实现高达660 TFLOPS的性能。)

2026/1/23

阅读全文 →

DeepSeek-V3.1混合推理架构解析：开启智能体时代新篇章

DeepSeek-V3.1 introduces hybrid inference with Think/Non-Think modes, enhancing reasoning efficiency and agent capabilities while supporting 128K context and updated APIs. (DeepSeek-V3.1引入思考/非思考混合推理模式，提升推理效率和智能体能力，支持128K上下文和更新的API。)

2026/1/22

阅读全文 →

DeepSeek R1代码优化能力解析：生成99% WASM性能改进代码

DeepSeek R1 demonstrates advanced code optimization capabilities, generating 99% of WASM performance improvements and showing superior reasoning in architectural decisions compared to other models. (DeepSeek R1展示了先进的代码优化能力，生成了WASM性能改进的99%代码，并在架构决策方面表现出优于其他模型的推理能力。)

2026/1/22

阅读全文 →

1 2 3...5

1 / 5