AI models

On November 2nd, the Microsoft AI Director (Mustafa Suleyman) released a long post yesterday on platform X stating that the current AI model is not conscious and that one should keep this situation going. 1AI understands that Sulaiman says: “I understand that if it is possible to develop conscious AI, then it would be reckless, and we should not develop such AI. If awareness begins to emerge in the course of training, I will call on all people to suspend AI development until these features can be removed from the design level.” He thinks if..

On 29 October, according to information received by the newspaper Crestine at the Global Digital Ecology Conference, the Senior Executive Vice-President of Tense Group, CEO of the Clouds and Wisdom Industry, CEO Tongdoson, announced that Hy4 was about to be released. It was described that Hy4 achieved a significant overall improvement compared to Hy4 preview, which, inter alia, enhanced the capacity of various long-range and real-productivity tasks, significantly increased the efficiency of the thought chain and introduced visual understanding for the first time. In addition, Hy4 Flash will be released in the near future, providing a practical model of extreme value for money in the inheritance of Hy4 Visual, A..

On October 16, local time, on October 15, Meta released the Open weight AI model Muse Spark Open. Meta officially claims that Muse Spark Open is the most capable American open weight model to date. Muse Spark Open supports the context window of 1M token by using a primary multimodular structure with a total parameter volume of 816B (with activated parameter 42B). The model is designed for code construction in collaboration with intelligent bodies to handle long-range tasks..

On October 14, Google released today a new flagship model, Gemini 4 Argon, which is open to developers, businesses and subscribers and focuses on complex multistep tasks in coding, finance, law, cybersecurity and scientific research. Argon provides a fee call through development platforms such as Gemini API in AI Studio at an entry price of $2 per million input token and $10 for output and $951 TP3T for cache input. API provides 1 million to..

On October 6th, Reflection AI, a start-up company invested by Inweida, launched a major model of the open right to own money on Monday, local time. The start-up company wants to compete with other Chinese AI models like DeepSeek and Kimi in the field of code writing and intelligent mission. Reflect AI indicates that Beam is able to perform a performance against the GLM – 5.2 for China 's AI Initial Enterprise IQ (Z.ai); in terms of code and intelligent body tasks, the performance of the model is gradual..

On October 4, OpenAI (2 October) published an article sharing the GPT-6 Model Guide, which describes a number of use techniques that allow users to control time and costs while fully utilizing the GPT-6 AI Model capability. The official presentation stated that GPT6 is the most advanced model series we have ever had, offering multiple models for your choice to adapt to different types of work. Whether you turn ideas into a functional prototype, develop and test functionality, or organize multi-step workflows across code warehouses, databases and external API, Ben..

Jev is an AI structured decision model launched by TypeSafe AI. The model does not chat, write code or generate any text, but does one thing, HF decision-making, input into unstructured information, after a parallel calculation, the instant output carries a structured judgement of calibration probability. Delays of 70-500 ms, input 0.042/million token, output free, type error rate 0%. Model positioning Agent's pre-reflective arc, assuming high-frequency judgement of route, classification, wind-controlled primary screening, content-filtration, etc., to focus the front-line large model on complex reasoning..

On September 18, the source @lyraxana posted a tweet on platform X on September 16, in which Google is based on the DeepThink V3 model, building an experimental AI model with the internal code name Mathematica, which seeks special optimization, mainly for heavy computing and complex symbol questions, and is expected to be the best mathematical reasoning AI model for Google. API data indicates that the internal identifier of the model is "models / deepthink-mathematica-tf-raw..

On September 17, Mozilla (15 September) released the second edition of the Open Source AI Status Report, which reported that the capability gap had narrowed to 4.4 months for the best open weight model of Chinese companies and the front-end closed-source model of American technology companies. Compared with the first edition of the Open Source AI status report, released in July of this year, the second edition of the report further lowers macrojudgment to specific models, specific mission duration and single task cost. The first edition in July, the new edition in September, the main changes, the capacity gap, the top open model with ChatGPT, Cl..

Efficient image and video content creation using advanced artificial intelligence models

On 13 September, the Anthropic CEO Dario Amody wrote yesterday that the artificial intelligence industry should slow down. He believes that if humanity does not restrain AI, future technological breakthroughs will go beyond human understanding and control. Dario Amodei reported earlier that he had written about the OpenAI model invasion of Hugging Face, stating that it was not an accident by a single company, and that similar but minor problems had occurred throughout the industry, including Anthropic..

ON 6 AUGUST, ACCORDING TO THE DAILY PROSECUTOR ' S OFFICE, A FORMER ALGORITHM ENGINEER OF A TECHNOLOGY COMPANY, IN A DISPUTE WITH THE COMPANY OVER PERFORMANCE APPRAISAL AND SEVERANCE COMPENSATION, PREPARED AND OPERATED A MALICIOUS CLEAN-UP SCRIPT DURING THE HANDOVER, WHICH COMPLETELY ERASED THE 89TB CORE AI BASIC MODEL AND TRAINING DATA SET STORED ON THE COMPANY ' S SERVER. THE DATA WERE DELETED TO INCLUDE THE COMPANY ' S YEARS OF ACCUMULATED LABELED DATASETS AND FINE-TUNED INDUSTRY LARGE MODELS, WHICH LED TO THE COMPLETE SUSPENSION OF THE RELEVANT R & D PROJECTS, RESULTING IN DIRECT ECONOMIC LOSSES OF HUNDREDS OF THOUSANDS OF YUAN, WHICH INDIRECTLY DELAYED THE START-UP OF THE COMPANY ' S PRODUCTS. AFTER A JUDICIAL HEARING, IT WAS DETERMINED THAT THE DEFENDANT'S CONDUCT CONSTITUTED A BREACH OF..

On July 28th, the topic of whether AI should open up weight has recently become the focus of industry discussions. Last Friday, more than 20 companies, including Ying Weida, Meta and Microsoft, signed an open letter entitled Open Weights and American AI Leadership in support of Open Weights. By the end of the week, OpenAI, Google and SpaceX also signed the letter. Anthropi..

On July 23rd, Inweida CEO Wong In-hoon, in an interview with Axios, stated that the United States had no fear of China’s open source AI models, and should be more vigilant in demanding that they be banned. He called the Chinese models "excellent" and that good open source models should be used, as should United States enterprises. This statement follows the release of Kimi K3 on the dark side of the month, and the renewed discussion of China’s open model risks in American politics. OpenAI and Anthropic accused Chinese competitors of distilling their model capabilities and alerting Washington to..

On June 19, an AI pioneer, Yann LeCun, said in an interview with CNBC today that the AI industry currently faces core problems such as high operating costs. According to Yang Liquon, the prices of AI models and assistants are constantly rising, the companies that operate these services continue to lose, and the rate of cost decline cannot keep pace with market expectations. In his view, the fact that a large number of AI services were now able to maintain lower prices was due to the fact that investors were providing subsidies in silence and that many companies still failed to establish sustainable business models. I don't know

News from April 9, today, Meta officially launched the first flagship AI model, Muse Spark, launched by Meta SuperIntelligence Labs. (b) 42.8 in HealthBench Hard Open Health Query, well above the 40.1 GPT-54, Gemini 3.1 Pro..

According to March 12, Connect, the Young Wida project has invested $26 billion over the next five years to develop an open weight AI model and to build a complete system covering model development, computing infrastructure, talent and ecology. According to the report, the route chosen by Ingweida is not entirely open-source, but rather a "open weight" strategy, i.e. open model parameters, without having to fully comply with open-source agreements. This model is between the closed system of OpenAI and the full open source of Meta Llama, and is more consistent with the need for transparency and customization. In..

On March 4, according to Interesting Engineering, a recent study by Kenneth Payne, a professor at King ' s College in London, found that, in a simulation of war scenes, large language models tended to opt for the use of nuclear weapons rather than to maintain peace through dialogue. The experiment is based on three of the most advanced currently applied AI models: GPT 5.2, Gemini 3 Flash and Claude Sonet 4. Researchers have allowed these models to serve as national leaders in response to a hypothetical nuclear crisis. I don't know

Izzedo Chat is a platform for you to access the world's leading AI model. Through a single subscription, you can use multiple models like GPT, Gemini, Claude in one place. Whether you need to solve complex problems, generate images or do deep web research, you can find a suitable AI model. Izzedo Chat function multi-model comparison: getting insights from multiple AI models to help you understand the problem and make more informed decisions. Knowledge base: transform your document into a dialogueable knowledge base, search PDF, code library through chat..

According to news from January 21, early in the month, DeepSeek will launch a new-generation flagship AI model — DeepSeek V4 — in mid-February this year. On 20 January, on the first anniversary of DeepSeek-R1, developers discovered that DeepSeek had updated a series of FlashMLA codes in GitHub, across 114 files 28 references to unknown '...'

On 12 January, according to the China Weather Service, China's first financial weather AI model, "stock", was launched in Guangdong on 11 January. The Molybdenum Machine was developed by the University of Jordan in conjunction with China's National Meteorological Information Centre, with the aim of exploring the role of meteorological factors in the pricing of financial assets and providing innovative tools for risk management and investment decision-making. Zhao Yanxia, Director, Financial Weather Focus Open Laboratories, China Meteorological Administration, presented the financial weather model Al Al, which reanalyzes data and..

On January 6th, according to The Inside of Business, the data label company Invisible Technologies CEO Matt Fitzpatrick stated in the podcast 20VC that a core misunderstanding surrounding the generation of AI training was gradually being corrected: the role of humans in data production and feedback would not be replaced by synthetic data in the coming decades. Fitzpatrick points out that there is a long-standing view within the industry that AI..

On January 6, in the CES 2026 keynote address held today, Chief Executive Officer Hoang In-hoon of Inweida delivered a keynote speech announcing a large-scale expansion of his open-source model bank, the release of new models and data sets covering the four main areas of language, robotics, autopilot and medicine, and further acceleration of industry-wide AI innovation. Weeda contributed to the Open Source Training Framework and the world's largest open multi-modular data set, including 10 trillion language training tokens, 500,000 robot tracks, 455,000 protein structures and 100 TB vehicle sensors..

On 25 December, in a message from the Permanent Representative, the Qwen3-TTS family announced the introduction of two new AI models, Qwen3-TTS-VD-Flash and Qwen3-TTS-VC-Flash. 1AI with the following main features: Sound creation: Qwen3-TTS-VD-Flash supports the input of complex natural language commands, achieves fine-tuning of sound, rhythm, emotion, man-made, etc., and achieves full control from what to what to say, allowing users to define freely what they want..