DeepSeek
S · ឈានមុខគេDeepSeek is an artificial intelligence company based in Hangzhou, China, and it receives investment from the hedge fund High-Flyer. It develops large language models that it has released publicly so anyone can use them. It also runs a service where people can chat with DeepSeek and offers an API that other services can connect to and use.
- #foundation-model
- #reasoning
- #open-weights
- #china
- of 108 AI companies and institutions
- #1
- សន្ទស្សន៍ LastAI
- 90.0
- ការវាយតម្លៃរបស់អ្នកប្រើប្រាស់
- —
- ការបោះឆ្នោត
- 0
- ចំនួនមើល
- 3
បច្ចេកវិទ្យា និងផលិតផលដែលពាក់ព័ន្ធ
វាយតម្លៃវាដោយខ្លួនអ្នក
ការបោះឆ្នោត 0 · ★ 0.0 (0)
ចូលដើម្បីបោះឆ្នោតនិងផ្តល់មតិយោបល់បច្ចុប្បន្នភាព
Tencent has released Hy4 Flash, a low-cost multimodal model that can handle up to one million tokens of context at once. The source article is a short summary and gives few details beyond the model's basic positioning.
A Korean article discusses how Chinese AI has long been labeled 'good enough' (cheap and usable but not the best). It argues this framing hides an assumption that frontier AI is judged by how well a single answer is given.

A Korean article discusses how Chinese AI has long been labeled 'good enough' (cheap and usable but not the best). It argues this framing hides an assumption that frontier AI is judged by how well a single answer is given.
A developer built a version of DeepSeek V3, a large language model, in about 200 lines of pure JAX code. The post was shared on GeekNews as a Show HN project.
A developer built a version of DeepSeek V3, a large language model, in about 200 lines of pure JAX code. The post was shared on GeekNews as a Show HN project.
A new web app lets people estimate how fast AI models will run on a local machine before buying hardware. The developer says users enter their hardware specs and model, and the app calculates how quickly it reads and writes text.

Public AI models from several Chinese companies show signs they were set up on Nvidia's Blackwell GPUs, which are restricted from export to China. Researchers found the traces by analyzing number patterns inside the models. The study does not prove Blackwell chips were used for the whole training process.

US intelligence reportedly concluded that Alice Hwang, once close to former Nvidia CEO Jensen Huang, was recruited by China's Ministry of State Security to gather information. Reports also show her associates invested in Chinese AI firm DeepSeek and had business ties with Huawei.
Public weights from six open models by DeepSeek, Moonshot AI, MiniMax, and Ant Group show initialization patterns matching Nvidia's Blackwell GPU. The analysis used numbers left in publicly released model files. It does not prove that all training ran on Blackwell chips.

DeepSeek, a Chinese AI company, reportedly took investments from firms linked to Alice Huang's family while it was getting access to Nvidia AI chips. The Information reported the story, citing anonymous sources. The investments came through special-purpose vehicles, so the related companies' names did not appear on DeepSeek's official shareholder register.

Companies linked to Alice Hwang, former partner of Nvidia's ex-CEO Jensen Huang, invested in Chinese AI company DeepSeek, according to The Information. The move was reportedly aimed at securing Nvidia's Blackwell chips, which are banned from sale in China.

Companies linked to Alice Huang reportedly invested in Chinese AI company DeepSeek and promised to supply it with Nvidia chips. The report came from the US outlet The Information on the 12th, local time.

Chinese AI company DeepSeek's autonomous agent framework, Harness, is growing its user base quickly. Harness is foundational software that helps AI work with files and run programs directly. The company plans to use this momentum to begin monetization through paid services.
DeepSeek Harness processes about 92 trillion tokens a day and has reached 1.54 million daily active users. Tokens are the small units of text that AI uses when reading and writing. A harness is a framework that helps AI agents finish tasks and connect to tools. The article says competition in the AI agent market has entered a full-scale phase.

DeepSeek's AI agent framework, Harness, has passed 1.54 million daily active users. About 92 trillion tokens (units of text that AI processes) are used each day. This is less than three months after launch.

DeepSeek's AI agent platform, DeepSeek Harness, became a service used by 1.54 million people a day just three months after its launch. External developers have built about 35,000 plugins for it. Based on this growth, DeepSeek plans to launch an official marketplace and expand support for developers.
The daily users of DeepSeek, a Chinese AI company, have passed 1.54 million. On the same day, it was reported that Nvidia's chief executive officer (CEO) is under investigation. You can check the full details in the original article.

Between late September and early October 2026, hacking incidents were confirmed at seven domestic financial companies. Investigators found signs that ARTEX, a program that lets AI automatically run security tools, was used in the attacks. The report says AI has started to take a direct part in carrying out attacks, not just serving as a simple tool.
llama-modchip is an open-source project that runs the DeepSeek-V4-Flash model on a computer with 64GB of memory and no graphics card (GPU). It compresses the model file and optimizes CPU calculations to make this possible. According to the project's own measurements, actual memory use in a 256K-token conversation was about 57.61 to 58.9GiB, and it did not use swap, which treats disk space as memory. These results are measured by the project itself.
llama-modchip is an open-source project that runs the DeepSeek-V4-Flash model on a computer with 64GB of memory and no graphics card (GPU). It compresses the model file and optimizes CPU calculations to make this possible. According to the project's own measurements, actual memory use in a 256K-token conversation was about 57.61 to 58.9GiB, and it did not use swap, which treats disk space as memory. These results are measured by the project itself.
អ្វីដែលអ្នកប្រើប្រាស់និយាយ
មិនទាន់មានមតិយោបល់ទេ។
ភស្តុតាងនៅពីក្រោយពិន្ទុនេះ
ភស្តុតាងរឹងមាំ- Real API usageទម្ងន់ 40
30.0% of all tokens · deepseek ·
Check on OpenRouter - Flagship productទម្ងន់ 25
Flagship index 74.1
ការដកស្រង់និងផ្លាកសញ្ញា
កូដយោង និងផ្លាកសញ្ញា
LastAI · v1.3 · · សំណុំទិន្នន័យ Snapshot
DeepSeek · ចំណាត់ថ្នាក់ក្រុម #1 · សន្ទស្សន៍ LastAI 90.0 (2026-11-20) · https://lastai.org/rankings/index.json?date=2026-11-20
ជ្រើសរើសអត្ថបទដើម្បីចម្លងវា។
<a href="https://lastai.org/en/rankings/deepseek-company"><img src="https://lastai.org/badge/deepseek-company.svg?lang=en" alt="DeepSeek" height="20"></a>
ជ្រើសរើសអត្ថបទដើម្បីចម្លងវា។
[](https://lastai.org/en/rankings/deepseek-company)
ជ្រើសរើសអត្ថបទដើម្បីចម្លងវា។
