DeepSeek V4
PulseAugur coverage of DeepSeek V4 — every cluster mentioning DeepSeek V4 across labs, papers, and developer communities, ranked by signal.
- developed by DeepSeek 100%
- subsidiary of DeepSeek 100%
- used by 36Kr 90%
- used by Huawei Ascend 90%
- used by Moore Threads 90%
- developed by DeepSeek V4-Flash 90%
- instance of DeepSeek V4-Flash 90%
- founded Liang Wenfeng 90%
- competes with Moonshot AI 80%
- used by SGLang 80%
- instance of DeepSeek 70%
- used by GitHub 70%
- 2026-05-25 research_milestone DeepSeek V4 completes full adaptation to Huawei Ascend chips, marking a milestone for China's domestic AI stack. 来源
- 2026-05-24 research_milestone DeepSeek V4 is presented as a new AI model challenging OpenAI.
- 2026-05-23 product_launch DeepSeek released its V4 model with a 1 million token context window. 来源
- 2026-05-18 product_launch DeepSeek launched its V4 models, V4-Pro and V4-Flash, on April 24, 2026.
- 2026-05-16 product_launch DeepSeek V4, a new open-weight LLM family, was released with significant architectural improvements and cost reductions. 来源
- 2026-05-16 research_milestone DeepSeek V4 achieves a 98% reduction in KV-cache memory usage with its new compressed attention architecture. 来源
- 2026-05-15 product_launch DeepSeek released its V4 model with MegaMoE optimizations. 来源
- 2026-05-11 product_launch DeepSeek V4, an AI model with a sparse Mixture-of-Experts architecture, was released.
- 2026-05-11 product_launch DeepSeek officially released its new flagship model, DeepSeek-V4. 来源
13 天有情绪数据
-
LLM Architectures Innovate for Long-Context Efficiency
Sebastian Raschka's analysis highlights recent architectural innovations in open-weight LLMs aimed at improving long-context efficiency. Key developments include KV sharing and per-layer embeddings in Google's Gemma 4 m…
-
DeepSeek V4 paper details algorithmic shifts in MoE scaling
DeepSeek V4, a new frontier model, has been detailed in a technical paper, showcasing significant advancements in Mixture-of-Experts (MoE) scaling. The paper delves into the algorithmic shifts that enable this scaling, …
-
NetEase News integrates DeepSeek-V4 to upgrade AI-powered services
NetEase News and NetEase Xiaomifeng have integrated the DeepSeek-V4 large language model to enhance their services. This integration aims to upgrade core functions such as news distribution, content creation, community …
-
DeepSeek V4 debuts with MegaMoE optimizations for efficient MoE
DeepSeek has released its V4 model, featuring significant optimizations through a new system called MegaMoE. This system utilizes a 1400-line fused CUDA kernel to enhance performance by fine-grained pipelining of commun…
-
Moore Threads rallies open-source AI dev community for MUSA GPU ecosystem
Chinese GPU maker Moore Threads has convened a meetup focused on integrating its MUSA architecture with key open-source large model inference frameworks like SGLang. The event brought together core developers from proje…
-
AI agents need web access, not just reasoning, to succeed
A developer with extensive OS and AI hardware experience argues that most AI agents fail due to their inability to bypass web security measures like Cloudflare. He introduces the concept of a "Full-Auto Browser Proxy" a…
-
NetEase Cloud Music integrates DeepSeek-V4 for enhanced user experience
NetEase Cloud Music has integrated DeepSeek-V4 to enhance user experience across various features, including music discovery and community interactions. This move signifies a broader trend of companies adopting advanced…
-
Ten new LLMs including DeepSeek V4, Grok 4.20, GPT-5.5 Pro to be benchmarked
A new benchmark test is scheduled to evaluate ten previously untested large language models, including DeepSeek V4 Pro, Grok 4.20, and GPT-5.5 Pro. The tests will focus on real-world agent coding tasks using a consisten…
-
Tencent Hunyuan 3.0 preview released with major architecture overhaul
Tencent's AI lab has released a preview version of its new model, Hunyuan 3.0, which marks a significant architectural overhaul focused on foundational elements. Led by Yao Shunyu, the team has prioritized data quality …
-
AI platform Lingzhu integrates DeepSeek V4, speeds up demand analysis
The AI creation platform Lingzhu has launched its second internal beta, featuring significant upgrades. Users can now access the platform without an invitation code and experience a notable performance boost due to the …
-
AI创作平台灵珠开启二测，接入DeepSeek V4模型
AI creation platform Lingzhu has launched its second internal beta test, removing the need for invitation codes and fully integrating the DeepSeek V4 large model. This integration significantly boosts efficiency, reduci…
-
DeepSeek-V4 launches with 1M context, Chinese hardware optimization
DeepSeek has officially released its latest flagship model, DeepSeek-V4, featuring a 1 million token context window and enhanced agent capabilities. The model comes in two versions, Pro and Flash, with the Pro version s…
-
Chinese AI industry focuses on domestic chips and LLMs
Dongwu Securities reports that the AI industry is undergoing a significant shift driven by intelligent agents and a push for domestic technological self-sufficiency. The training of DeepSeek V4 using domestic computing …
-
DeepSeek-V4's 1M-token context window is an inference systems challenge
Together AI has detailed the architectural innovations behind DeepSeek-V4's ability to handle a 1 million token context window. The model employs a hybrid attention design that compresses context before storing it in th…
-
mlx-audio adds TTS models, AMD ROCm sees performance boost, Claude Code architecture detailed
mlx-audio v0.4.3 has been released, introducing six new text-to-speech models and optimizing dependencies for Apple Silicon developers. Separately, AMD's ROCm software stack saw a significant 75x performance increase wi…
-
DeepSeek V4 benchmarks show 85 tok/s at 524k context; Ollama guide for Ryzen APUs released
New benchmarks reveal DeepSeek V4 Flash achieving 85 tokens per second with a 524k context window, utilizing MTP self-speculation and FP8 quantization on dual RTX PRO 6000 Max-Q GPUs. Additionally, a guide has been publ…
-
Alibaba's open-source AI models lead in adoption
Alibaba's open-source AI models, DeepSeek-V4 and Qwen, have reportedly surpassed competitors in adoption rates. This achievement highlights China's growing influence in the open-source AI landscape.
-
OpenAI's GPT-5.5 prioritizes reliability for production AI agents over benchmarks
OpenAI has released GPT-5.5, which reportedly excels not in benchmark scores but in practical reliability for complex tasks. The new model demonstrates significantly improved instruction following, reduced hallucination…
-
Xiaomi Auto reshuffles leadership as Nvidia eyes optical expansion for AI
Nvidia CEO Jensen Huang praised a new partnership with Corning, highlighting its potential to bolster the US tech supply chain. Huang emphasized the growing need for optical connectivity in next-generation AI infrastruc…
-
DeepSeek-V4 powers GitHub trending Agent project, while Xiaomi faces sales dip
A new investment fund, Lanzhou Future Innovation Industry Development Fund, has been established with approximately 200 million RMB in registered capital. The fund, managed by Gansu Xinglong Fund Management Co., Ltd., w…