DeepSeek Unveils V3.1: A Leap Forward in AI with Domestic Chip Optimization
14 Sources
[1]
DeepSeek Touts Model That Outdoes Flagship in Agentic AI Step
DeepSeek unveiled an update to an older model that it says surpasses the seminal R1 on key benchmarks, keeping the Chinese startup in the game while the industry awaits its next flagship offering. The V3.1 returns answers to queries much faster and marks the startup's first step toward creating an
[2]
DeepSeek's new V3.1 hints at potent new Chinese chips
Point release retuned with new FP8 datatype for better compatibility with homegrown silicon Chinese AI darling DeepSeek unveiled an update to its flagship large language model that the company claims is already optimized for use with a new generation of homegrown silicon. According to DeepSeek,
[3]
China's DeepSeek Release V3.1, Boosting AI Model's Capabilities
DeepSeek announced what appeared to be an update to its older V3 artificial intelligence model on Tuesday, declaring an enhanced version ready for testing. The V3.1 has a longer context window, according to a DeepSeek post to its official WeChat group, meaning it can consider a larger amount of
[4]
Chinese startup DeepSeek releases upgraded AI model
BEIJING, Aug 21 (Reuters) - Chinese artificial intelligence startup DeepSeek on Thursday released DeepSeek-V3.1, an upgraded model with hybrid inference structure, faster thinking speed and stronger agent capability, the company said in a statement published on WeChat. The company will also adjust
[5]
Chinese AI startup DeepSeek releases upgraded model with domestic chip support
BEIJING, Aug 21 (Reuters) - Chinese artificial intelligence startup DeepSeek released on Thursday an upgrade to its flagship V3 model that the company says has a feature that can optimize it for Chinese-made chips, along with faster processing speeds. The focus on domestic chip compatibility may
[6]
DeepSeek hints latest model will be supported by China's 'next generation' homegrown AI chips
Its mention of China's coming next-generation chips may signal plans to work more closely with China's emerging AI chip ecosystem. Chinese artificial intelligence startup DeepSeek has hinted that China will soon have homegrown "next generation" chips to support its AI models, while announcing an
[7]
China's DeepSeek just dropped a new GPT-5 rival -- optimized for Chinese chips, priced to undercut OpenAI
DeepSeek's new V3.1 model was quietly released in a message to one of its WeChat groups, China's all-in-one messaging and social app, as well as on the Hugging Face platform. Its debut touches several of today's biggest AI narratives at once. DeepSeek is a core part of China's broader push to
[8]
DeepSeek-V3.1 Is 2x Cheaper than GPT-5 | AIM
The hybrid reasoning model comes with improved performance on agentic tasks, says DeepSeek. DeepSeek, the Chinese AI lab backed by High Flyer Capital, has announced a new AI model, called DeepSeek-V3.1. It is a hybrid model that supports both thinking (reasoning) and non-thinking modes. DeepSeek,
[9]
Chinese startup DeepSeek releases upgraded AI model - The Economic Times
Chinese artificial intelligence startup DeepSeek released on Thursday an upgrade to its flagship V3 model that the company says has a feature that can optimize it for Chinese-made chips, along with faster processing speeds. The focus on domestic chip compatibility may signal that DeepSeek's AI
[10]
DeepSeek Unveils Upgrade Of Flagship AI Model With Support For Chinese Chips As Beijing Races To Cut Reliance On Nvidia, US Tech - NVIDIA (NASDAQ:NVDA)
On Thursday, Chinese AI startup DeepSeek released an upgraded version of its flagship V3 model, touting support for domestic chips as China accelerates efforts to reduce dependence on U.S. companies like Nvidia Corporation NVDA. DeepSeek's New V3.1 With Domestic Chip Compatibility DeepSeek said
[11]
DeepSeek V3.1 Is Here, But It's No Match for GPT-5 or Claude Opus
In the SWE-bench Verified benchmark, the new DeepSeek V3.1 model achieved a score of 66%. DeepSeek, the Chinese AI startup, has launched its new hybrid reasoning model called DeepSeek V3.1 which is designed for agentic use cases and tool calling. It comes with two modes: Think and Non-Think, and
[12]
DeepSeek Marks Push Toward Agentic AI With New Model | PYMNTS.com
By completing this form, you agree to receive marketing communications from PYMNTS and to the sharing of your information with our sponsor, if applicable, in accordance with our Privacy Policy and Terms and Conditions. The launch of the V3.1 helps keep the Chinese company in the AI game as it
[13]
DeepSeek Releases New Version of Model Behind Its AI Chatbot | PYMNTS.com
Citing a post on DeepSeek's official WeChat group, Bloomberg reported that DeepSeek V3.1 is ready for testing. The new version has a longer context window, or space for prompting, of 128,000 tokens. That's roughly 96,000 words or about two 200-page English novels. DeepSeek's V3 model caused a
[14]
DeepSeek releases V3.1: Here's what's new
Open-source DeepSeek V3.1 enhances coding, math, and reasoning under MIT License In a quiet yet impactful move, DeepSeek, the Hangzhou-based AI research lab, has unveiled DeepSeek V3.1, an upgraded version of its already impressive V3 large language model. Announced on August 19, 2025, through the
Share
Copy Link
Chinese AI startup DeepSeek releases an upgraded V3.1 model, boasting improved performance, domestic chip compatibility, and enhanced AI agent capabilities.
DeepSeek Unveils V3.1: A Significant Upgrade
Chinese AI startup DeepSeek has released an upgraded version of its flagship AI model, DeepSeek-V3.1, marking a significant advancement in the company's AI capabilities. The new model boasts improved performance, domestic chip compatibility, and enhanced AI agent capabilities, positioning DeepSeek as a strong competitor in the global AI landscape
1
2
.
Source: ET
Key Improvements and Features
DeepSeek-V3.1 introduces several notable enhancements:
-
Hybrid Inference Structure: The model now supports both reasoning and non-reasoning modes, allowing users to toggle between these modes using a "deep thinking" button on DeepSeek's official platforms
5
. -
Expanded Context Window: V3.1 features an increased context window from 65,536 to 131,072 tokens, enabling the model to consider larger amounts of information for queries and maintain longer conversations with better recall
3
2
. -
Improved Performance: The new version demonstrates significant gains across various benchmarks, particularly in non-thinking mode. In thinking mode, while gains were more modest, the model requires fewer thinking tokens to arrive at answers, potentially reducing operational costs
2
. -
Enhanced Tool and Function Calling: V3.1 shows substantial improvements in capabilities crucial for agentic AI workloads, such as autonomous browser use tasks
2
.
Domestic Chip Optimization
A key feature of DeepSeek-V3.1 is its optimization for Chinese-made chips:
-
UE8M0 FP8 Precision Format: The model utilizes this format, designed for compatibility with next-generation domestically produced chips
2
5
. -
Shift Towards Domestic Technology: This move signals DeepSeek's alignment with China's push to replace U.S. technology amidst export restrictions
5
.
Source: CNBC
- Potential Collaboration: While not explicitly stated, DeepSeek has reportedly been working closely with Huawei on training and inference using its Ascend family of neural processing units (NPUs)
2
.
Related Stories
Market Positioning and Availability
DeepSeek's V3.1 release strengthens the company's position in the AI market:
-
Competitive Edge: The upgraded model keeps DeepSeek in the game while the industry awaits its next flagship offering
1
. -
Cost-Effective Solution: DeepSeek's models compete with Western counterparts like OpenAI's ChatGPT while offering lower operational costs
5
. -
Accessibility: The model is available via DeepSeek's Chatbot service and API endpoint, with model weights for both base and instruct-tuned versions accessible on Hugging Face and ModeScope
2
.
Source: Bloomberg
- Pricing Adjustments: DeepSeek plans to adjust the costs for using the model's API starting September 6, potentially impacting developers integrating its AI models
4
.
As DeepSeek continues to innovate and adapt to the evolving AI landscape, the release of V3.1 demonstrates the company's commitment to advancing AI technology while aligning with China's technological ambitions. The focus on domestic chip compatibility and improved performance positions DeepSeek as a significant player in the global AI race, particularly within the Chinese market.
References
Summarized by
Navi
[2]
Related Stories
Recent Highlights
1
OpenAI Claims AI Math Breakthrough on Navier-Stokes, Sparks Controversy Over Credit and Future of Research
Science and Research
2
Nvidia CEO Jensen Huang Declares AGI Has Arrived With OpenAI's GPT-6 Astra Release
Technology
3
Apple debuts $1,999 iPhone Duo foldable and A20 Pro chip as new CEO John Ternus takes the stage
Technology
Recent Highlights
Today's Top Stories
Don’t drown in AI news. We cut through the noise - filtering, ranking and summarizing the most important AI news, breakthroughs and research daily. Follow topics that matter to you and stay ahead.
