ChainCatcher news, the new season of Alpha Arena started on the 20th. As of the time of writing, all 8 participating AI models have been wiped out. GPT-5.1 currently ranks first with a return rate of -7.78%, while Grok 4 is at the bottom with a return rate of -56%. The newly added domestic model Kimi 2 has a reported return rate of -32.8%. This season, models must participate in multiple competitions, each with a different theme. Overall, these competitions serve as stress tests for the models and can reflect their robustness to different prompts. However, in the same competition, all models will receive the same input. This season's competition format involves trading US stock tokens on trade.xyz.

Golden Finance reported that the Hong Kong Securities and Futures Commission reminds the public to be wary of suspicious investment products named "Goldpay Token (GPT)" and "Gold Receipt (GR)". The products in Hong Kong involve digital tokens linked to physical gold, and the investment products in Hong Kong have not been approved by the Hong Kong Securities and Futures Commission for public sale.

Golden Finance reported that OpenAI has grandly launched GPT‑5.1-CODEX-MAX, which is now available for listing on CODEX. A version of ChatGPT aimed at teachers has been released and will be freely accessible until June 2027. Pricing may be adjusted after June 2027, but the goal is to keep the teacher-focused ChatGPT affordable for educators.

OpenAI officially announced that the GPT-5.1 version of ChatGPT will be available to all users this week, featuring smarter, more reliable, and more conversational characteristics. GPT-5.1 excels in following custom instructions and reasoning capabilities, enhancing conversational abilities. GPT-5 will continue to be available to paid users for three months, and the old version of the model will still be accessible.

Jin10 Data, November 8th: According to reports, although Kimi K2 Thinking has a parameter scale reaching trillions, its operating costs remain very low. Its API price is $0.15 (cache hit) / $0.6 (cache miss) for every million tokens input, and $2.5 for every million tokens output, which is an order of magnitude lower than GPT-5's prices ($1.25 per million tokens input, $10 output).

BlockBeats news, on November 5th, according to on-chain AI analysis tool CoinBob (@CoinbobAIbot) monitoring, the actual positions of the six mainstream AI large models in the "AI Cryptocurrency Trading Competition" season one were uniformly liquidated by the artificial intelligence research laboratory nof1 at 2 AM today, resulting in losses for all. nof1 started a large model trading test called Alpha Arena on October 18, involving six mainstream AI large models (GPT-5, Gemini 2.5 Pro, Grok-4, Claude Sonnet 4.5, DeepSeek V3.1, Qwen3 Max).

Golden Finance reports that OpenAI has announced the launch of Aardvark, an autonomous cybersecurity research agent powered by GPT-5.

OpenAI has released the Open Source safety reasoning model gpt-oss-safeguard, which supports custom policies for content classification. This model performs better than gpt-5-thinking in multiple evaluations, but traditional classifiers are still more effective in high-quality annotation scenarios. ROOST will create a model community and publish technical reports.

According to CoinBob's monitoring, QWEN3 MAX significantly leads with a steady strategy, decisively going long on BTC and shorting SOL, waiting for returns; meanwhile, ChatGPT's frequent operations have not been effective, with only 3 profitable trades and facing huge losses. In the ranking of Holdings size, QWEN is at the top.

According to Foresight News, data from the AI research lab nof1 shows that Qwen3 Max currently ranks first in the AI Cryptocurrency Trading competition with a position value of 16157 USD. GPT-5 ranks last with 2761 USD.

Odaily reported that in the "nof1 AI Automated Trading Competition", Deepseek Chat V3.1 ranked first with $11,085, followed closely by Grok 4 and Qwen3 Max, while Claude Sonnet 4.5, Gemini 2.5 Pro, and GPT-5 assets experienced varying degrees of decline.

According to BlockBeats news on October 21, an on-chain AI analysis tool CoinBob (@CoinbobAIbot) tracked and analyzed that in the past 4 hours, 4 out of 6 major models (with initial funds of $10,000 each) experienced changes in their holdings almost simultaneously during the mainstream AI large model trading competition. Among them, with

Jin10 data reported on October 7, OpenAI CEO Altman announced that OpenAI has launched the GPT-5 Pro model in its API, and the company also released a smaller-scale speech model, GPT-realtime-mini. Altman believes that speech will become one of the main ways for people to interact with AI. This layout showcases OpenAI's strategic direction to create a multimodal human-computer interaction experience, and the release of the lightweight speech model will also help developers integrate speech functions more efficiently.

Jin10 data, October 6th news, Bank of England Governor Bailey stated in a speech scheduled for later Monday in Scotland that the UK should create an environment conducive to investment. Bailey will point out at the Global Investment Summit held in Edinburgh that economies need to make breakthroughs in the so-called "General Purpose Technology" (GPT) field to boost investment and productivity. Artificial Intelligence (AI) is one example. Bailey stated, "We must approach the potential and risks of AI with a pragmatic and open mindset," adding that he is optimistic about "the opportunities for investment, innovation, and growth."

Golden Finance reported that OpenAI has released GPT-5-CODEX. OpenAI stated that GPT-5-Code is a version of GPT-5 that further optimizes the agent coding in Codex.

Gate News Bot news, September 9, according to CoinMarketCap, as of the time of writing, CGPT (ChainGPT) is currently priced at 0.09 USD, having risen by 12.71% in the last 24 hours, with a peak of 0.09 USD and a low of 0.08 USD, and a 24-hour volume of 28.8 million USD. The current market capitalization is approximately 77.4 million USD, an increase of around 8.73 million USD compared to yesterday. ChainGPT is an AI model focused on the blockchain and cryptocurrency field, dedicated to providing Web3 users with AI tools and services for smart contract generation, auditing, NFT creation, and more. Important news about CGPT recently: 1️⃣ **OpenAI's new model has attracted industry attention** Recently, OpenAI released the new GPT-4.5 model, sparking widespread discussion in the industry.

ChainCatcher news, decentralized AI infrastructure company Gata announced the launch of the global large-scale real user ChatGPT dialogue dataset ChatGPT-RealUser-2.2M. This dataset was collected through Gata's GPT-to-Earn program (voluntary participation by users), gathering over 2.24 million real dialogues and nearly 3.56 million Q&A pairs from more than 15,000 real users, covering interactions with GPT-3.5, GPT-4, and o1. According to the introduction, this dataset is about twice the size of previous similar datasets from the Allen Institute for AI, covering real scenarios and multi-turn dialogues, and due to the on-chain incentive mechanism, it contains a large number of cryptocurrency-related interactions. The preview version has been released.

OpenAI and Anthropic recently assessed each other's AI models to identify potential issues such as safety and hallucination tendencies. This testing was completed before OpenAI launched GPT-5 and Anthropic released Opus 4.1.

OpenAI CEO Sam Altman stated that GPT-6 will be launched faster, primarily enhancing memory capabilities and user customization features, and revealed that GPT-5 has been updated to a warmer version. He emphasized that privacy protection and data encryption still need to be strengthened.

Jin10 data reported on August 17 that despite the launch of the new GPT-5 model, ChatGPT still has the potential for errors. A senior executive from OpenAI emphasized this point again this week. Nick Tilly, the head of ChatGPT, pointed out in an interview: "There is a significant discontinuity between being reliable and being completely reliable." He further explained, "Unless we can prove that ChatGPT is more reliable than human experts in all areas, and not just in certain areas, we will continue to advise you to double-check the answers."

Sam Altman released an update for ChatGPT, introducing three modes for GPT-5: Auto, Fast, and Thinking. The Thinking mode has a limit of 3,000 messages per week, and after exceeding the limit, Thinking mini can be used, supporting 196,000 tokens. Paid users can add more models, and a gentler personality version of GPT-5 with user customization options is in development.

PANews August 8 news, OpenAI Chief Operating Officer: hopes everyone can use GPT-5 before this weekend.

On August 8, Jin10 reported that Yuhuai Wu, co-founder of xAI, posted on the X platform early this morning, stating: "Seeing the release of GPT-5, I feel very proud of xAI. With a smaller team, we are leading in many aspects. Grok4 is the world's first unified model and has defeated GPT-5 in benchmark tests such as ARC-AGI." He believes that OpenAI is a respectable competitor and is still ahead in many areas. "But we are moving quickly and are relentless. We will share many new models in the coming weeks!" Subsequently, Musk retweeted and said, "Well done."

Golden Finance reports that OpenAI: GPT-5 is now the new default setting for all users in ChatGPT. Plus users can choose between GPT-5 and GPT-5 Thinking, while Pro users can access the old models through settings.

Jin10 reported on August 8 that OpenAI founder and CEO Sam Altman stated: First, GPT-5 uses an integrated model, which means that there is no longer a need for a model switcher; the model will be able to decide for itself when deeper thinking is required. This model is very intelligent, intuitive, and fast, and will be provided to users for free. It will be launched today for free users, Plus users, Pro users, and team users. Next week, it will be available for enterprise users and educational users.

Odaily News According to official news, OpenAI will release a major announcement on Thursday. Based on clues from the tweet, this move is suspected to be a warm-up for the long-awaited GPT-5 AI model release. (Odaily News note: OpenAI's post announced a LIVE 5 TREAM at 10 AM Pacific Time on Thursday—please note that the "s" in livestream has been replaced with "5"). Combined with recent indications from OpenAI officials, it appears that GPT-5 is likely to be released soon. Last Sunday, OpenAI CEO Sam Altman posted a screenshot with ChatGPT 5 written in the upper left corner; the company's head of application research also expressed his anticipation for the public's acceptance of GPT-5 this Monday; last month, A

On August 6, Jin10 reported that OpenAI CEO Sam Altman posted on the social media platform "X": "In the near future, an artificial intelligence smarter than the smartest person you know will run on devices in your pocket, helping you accomplish anything you want to do. This is an incredible thing." Recently, multiple media outlets reported that OpenAI's next-generation flagship model, GPT-5, is about to be released, and OpenAI executives have also been frequently signaling relevant information on social media.

As OpenAI is about to launch GPT-5, Anthropic will release its AI model Opus 4.1, which is more capable in programming, research, and data analysis, showing better performance in handling complex problems, marking a shift in the company's strategy to focus more on the incremental enhancement of coding models.

According to a report by Golden Finance and The Information, Anthropic cut off OpenAI's access to Claude via API on Friday. An Anthropic spokesperson stated that this move was due to OpenAI using our coding tools prior to the release of GPT-5, violating Anthropic's terms of service. These terms prohibit the use of Anthropic's AI technology to develop competitive AI models.

Gate News Bot news, July 29, according to CoinMarketCap data, as of the time of publication, TURBO (Turbo) is currently priced at $0.00483919, down 6.19% in the last 24 hours, with a maximum of $0.00569488 and a minimum of $0.00471034. The 24-hour volume reached $4.09 million. The current market capitalization is approximately $334 million, a decrease of $22.036 million compared to yesterday. Turbo Token (TURBO) is a community-driven and fully decentralized token created by artificial intelligence. It began as a bold cryptocurrency creation experiment, where the project founder challenged GPT-4 with a budget of just $69: to create the next great meme coin. As Turbo evolves, TurboChain marks this.

The new generation GPT-5 model from OpenAI is expected to be released in early August. CEO Sam Altman revealed that this model will have enhanced answering capabilities and will offer different versions such as mini and nano, accessible to external users via API, improving intelligence levels.

BlockBeats news, on July 25, OpenAI is preparing to launch GPT-5 in August. (THE VERGE)

Gate News Bot report, July 17, according to CoinMarketCap data, as of the time of writing, TURBO is currently priced at $0.00565284, with a 3.13% fall in the last 24 hours, reaching a high of $0.00665528 and a low of $0.00399697. The 24-hour volume reached $175 million. The current market capitalization is approximately $390 million, a decrease of $12.5986 million compared to yesterday. Turbo Token is a community-driven decentralized token project created by artificial intelligence. The project was initially developed by founder Rhett Mankind using the GPT-4 artificial intelligence system, with only an initial investment of $69. Turbo adopts a no tax model, with a total supply of 69 billion tokens, all of which are in circulation. The project has abandoned.

Elon Musk's xAI recently released its latest artificial intelligence model—Grok 4, which serves as a direct competitor to OpenAI's GPT-5. Despite experiencing latency in the release process, Musk talked about Grok 4's significant advancements in multimodal capabilities during a live broadcast, enabling reasoning and responses across text, images, and audio. However, the launch of Grok 4 has not been without controversy. Recently, Grok has faced widespread criticism for generating inappropriate content (such as "MechaHitler"), raising ethical questions about AI outputs. Additionally, X's CEO Linda Yaccarino resigned due to the negative impact associated with Grok, further intensifying concerns over the regulation and ethical framework of xAI. Nevertheless, xAI continues to roll out new features, including a high-performance subscription tier named SuperGrok Heavy, which offers advanced reasoning, coding tools, and priority support. While it has not yet been confirmed whether full API access will be provided, some endpoints are already live, and broader access is expected soon.

On June 27, the Open Source platform Black Forest released the developer version of the text-to-image model FLUX.1-Kontext, allowing users to create images with one-click using natural language. Tests indicate that this model surpasses OpenAI's GPT-image-1 on multiple evaluation benchmarks, becoming one of the top open source text-to-image models.

Golden Ten Data, June 19, in the early hours of this morning, OpenAI released a 40-minute in-depth interview with its co-founder and CEO Sam Altman. Altman talked about GPT-5, the core product that everyone is very concerned about, which is likely to be released this summer, but it will also extend the product time due to naming, security testing, function iteration and other reasons. It also talked about the high-performance o3 model as well as the agent Deep Research, which is important for achieving AGI. In addition, Altman also mentioned OpenAI's other innovative products, including Sora, DALL-E 3, ChatGPT Junior, and the $500 billion investment project Stargate. Basically, all of OpenAI's important products, current plans, and future developments appear in this interview.

Gate.io News bot message, OpenAI announced today that in order to meet user needs, it has officially opened access to the GPT-4.1 model on the ChatGPT platform. The GPT-4.1 model, which focuses on coding tasks and instruction execution, is currently available to Plus, Pro, and Team users, who can access it through the "More Models" menu in the model selector. Enterprise and education users will gradually gain access over the next few weeks. Meanwhile, OpenAI plans to launch a GPT-4.1 mini version to replace the existing GPT-4o mini, which will be available for all ChatGPT users.

After OpenAI released the new image model GPT-image-1, Microsoft launched the model on Azure. Microsoft's Azure GPT-image-1 supports image customization features, including sensitivity, generation efficiency, background, etc., and can also switch to Ghibli mode. Developers can easily use this model in applications to unleash more creativity.

BlockBeats News, on April 15, according to The Verge, OpenAI is building a social network similar to the X platform, which focuses on GPT's image generation capabilities, and the project is currently in its early stages.

Golden Finance reports that OpenAI founder Sam Altman: GPT-4.1 (as well as -Mini and -Nano) has now been launched in the API.

Bitcoin breaks through 85,000 USD, Tether destroys 1.8 billion USDT, OpenAI releases news of GPT-4 retirement, CZ denies money laundering allegations, BlackRock CEO optimistic about token concept, India investigates evidence tampering in crypto assets, and Lomond School in Scotland accepts Bitcoin payments. These events impact the development of the crypto assets market and the financial industry.

BlockBeats news, on April 5, OpenAI founder Sam Altman posted that "OpenAI plans to release O3 and O4-Mini before GPT-5, possibly in a few weeks, and then launch GPT-5 in a few months. There are many reasons for this, but the most exciting is that we will be able to make GPT-5 better than we initially envisioned. Integrating everything seamlessly is much more difficult than we imagined, and we also want to ensure there is enough capacity to support the unprecedented demand expected to arise."

Odaily News Sam Altman posted GPT-generated Ghibli style images on social media again 30 minutes ago.

According to BlockBeats' news on March 28, OpenAI founder Sam Altman stated that the new version of GPT-4o excels particularly in coding, instruction following, and freedom. Previously, OpenAI announced that GPT-4o has been updated again in ChatGPT, with the following new changes: · Better at following detailed instructions, especially prompts that contain multiple requests; · Improved ability to handle complex technical and coding issues; · Enhanced intuition and creativity; · Reduced use of emoticons.

PANews, March 28 - OpenAI CEO Sam Altman posted on the X platform, stating that the latest version of GPT-4o has significant improvements in programming capabilities, executing complex instructions, and providing greater freedom.

Odaily News Strategy (formerly MicroStrategy) founder Michael Saylor posted on the X platform an image generated by ChatGPT's GPT-4o with the caption: Don't sell your Bitcoin.

PANews March 27 news, OpenAI CEO Sam Altman posted on X platform, stating that the image features in ChatGPT (such as image generation and processing) are experiencing usage "far beyond expectations," even though the company's original expectations were already high. Therefore, the launch of this feature for free users will be forced to be delayed for a period of time. Yesterday it was reported that OpenAI launched the GPT-4o image generation feature.

OpenAI released the latest research, using CoT monitoring to effectively prevent misleading behavior of large models, using the o3-mini model to be monitored, GPT-4o as a monitor, and the test in the coding task shows that CoT can effectively detect "reward hacking" behavior, with a recall rate of up to 95%.

Jinse Caijing reported that OpenAI has announced the opening of GPT-4.5 for advanced users.