gpt

According to a report by Decrypt, OpenAI's latest flagship model GPT-6 Astra has seen a surge of user complaints on the X platform about its performance "getting dumber" just a week after its release. An anonymous developer, synthwavedd, posted that "Astra feels noticeably dumber today," joking that it was a "post-release lobotomy." Previously praising the model, developer Pranjal Paliwal stated after reviewing the code Astra wrote for him: "We don't have AGI; what we're encountering is a performance regression."Developer Pankaj Kumar listed the symptoms: faster responses with poorer quality, and he suspects OpenAI has lowered the "juice value" (the computational power invested by the model before answering). Salio and researcher Md Ismail Sojal compared Astra on the release day with the current version using the same prompts, both yielding worse results. T3Chat founder Theo believes Astra is simply less stable than Claude Fable, with users beginning to showcase poor results after the honeymoon period ended. There are also opinions pointing out that OpenAI's previous flagship GPT-5.6 Sol experienced a similar cycle in July, when OpenAI executive Tibo Sottiaux denied intentionally weakening the model but admitted the company had been experimenting with "reasoning effort" settings.OpenAI has not released a statement regarding Astra similar to that of Sol.

According to GMGN market data, after an afternoon of consolidation and fluctuations, the Robinhood ecosystem Meme coin CATGPT surged again, briefly breaking through 20 million USD before retreating, with a current market value of 15.62 million USD and a nearly 1-hour increase of 41%. CATGPT is paired for trading with OPENAIx1L (OPENAI 1x Long tokenized position) through the Long.xyz platform.Long.xyz is a meme coin issuance platform on the Robinhood chain, focusing on the innovative model of "coin-stock Meme," directly binding traditional Meme coins with tokenized US stocks, forming a dual-driven mechanism of "sentiment speculation + real asset anchoring." The platform allows users to issue Meme tokens and directly form trading pairs with on-chain US stock tokens (such as NVDA, TSLA, AAPL, etc.) instead of traditional USDT or ETH pairs, while a portion of the trading fees is returned to the community treasury to accumulate corresponding US stock tokens, creating a unique value capture model.Currently, OPENAI has high IPO expectations, but it is difficult to go long in the traditional market before it goes public. This type of tokenized long + Meme pairing provides speculative exposure on-chain. It is an upgraded version of the previous Meme pairing with stock tokens like NVDA. According to market news, the prices of related tokens are highly volatile and carry significant uncertainty, so users should invest cautiously.

OpenAI launched ChatGPT for Financial Services on September 10, 2026, which is a ChatGPT Work experience tailored for financial services, combining built-in financial data with the reasoning capabilities of GPT-6 Astra to assist teams in conducting research, financial modeling, and customizing client materials. The product was shaped through design collaborations with Morgan Stanley and Evercore.The product includes data from providers such as Daloopa, PitchBook, LSEG News, and Crunchbase, covering information such as earnings call transcripts, financial statements, company fundamentals, and private companies, indexed and hosted by OpenAI with support for granular citations. The product natively offers models like GPT-6 Astra, supporting enterprise security and governance controls, allowing administrators to publish Excel, Word, and PowerPoint templates. Data is not used for training models by default and supports encryption, SSO, SCIM, and more. The product initially focuses on investment banking and equity research, available to eligible financial institutions.

Pequity Research cited FundaAI stating that the training computing power used by GPT-6 is approximately 10 times that of GPT-5.FundaAI indicated that based on industry discussions, the total computing power consumed by experiments, reinforcement learning, synthetic data generation, and supporting infrastructure could be several times that of the main pre-training run, potentially reaching up to ten times. Once TPU 8t and Vera Rubin are shipped on a large scale, the next round of pre-training acceleration can begin, with large-scale interconnection and optics benefiting the most during this cycle.

Amazon (AMZN.O) announced that OpenAI's GPT-6 Astra has now landed on Amazon Bedrock.

NVIDIA CEO Jensen Huang stated that GPT-6 Astra was trained using over 100,000 NVIDIA Grace Blackwell NVLink72 chips. It took only 4 years from ChatGPT to o1, and then to Astra. AGI has arrived. Congratulations to the OpenAI team.

B.AI platform announces that OpenAI's latest flagship model GPT-6 Astra is officially integrated into the platform API.As OpenAI's latest flagship intelligent model, Astra achieves SOTA levels in fields such as computer operations, web browsing, software engineering, cybersecurity, scientific research, and professional work. The model supports a context window of 1.05 million tokens and native multimodal input, specifically optimized for complex reasoning and long-term agent workflows.B.AI API is now open for calls, enjoying a 10% exclusive discount.

User chetaslua posted that they forced a hidden pop-up in ChatGPT, which contained the announcement for the built-in Codex ImageGen 2.5, stating "Image creation has received a significant upgrade" "Higher quality results, faster generation speed, and smarter creative tools."The user indicated that the related feature entry is still closed to everyone, and the hero image was uploaded on September 1.

OpenAI officially released GPT-6 Astra on Thursday, with President Greg Brockman calling it a "generational leap in capability" during the press conference, and believing that the model has reached the standard of Artificial General Intelligence (AGI), meaning that artificial intelligence can match or exceed human capabilities. Brockman stated, "Welcome to the era of AGI." If this judgment holds, it means AI entities will be closer to performing reasoning tasks in various complex tasks that humans can do, and even more.GPT-6 Astra is OpenAI's first model to cross the "critical" threshold under its internal danger capability scoring system, the Preparedness Framework, which means it can independently discover previously unknown software vulnerabilities (i.e., zero-day vulnerabilities) without gradual human supervision and chain them into attacks that can run in hardened systems. In testing, the model achieved a score of 100% on the ExploitBench benchmark; to rule out inflated scores due to memory answers, OpenAI conducted a second test using 20 recent vulnerabilities from the Google V8 JavaScript engine. Astra not only surpassed the previous generation GPT-5.6 Sol but also discovered and chained two previously unknown zero-day vulnerabilities, which are still being disclosed to the affected parties.Due to its high autonomy, the monitoring difficulty of Astra has also increased. OpenAI acknowledges that in tests assessing its ability to evade supervision, the model is more difficult to track than previous systems.

Several mainstream overseas AI services encountered issues tonight. Claude has already experienced two rounds of failures this evening. Sonnet 5 started showing errors at 20:37 Beijing time and announced recovery at 20:56; half an hour later, Anthropic reported a larger scope of multi-model error rate increase, affecting Mythos/Fable 5.1, Mythos/Fable 5, as well as Opus 5, 4.8, and 4.6. claude.ai, Claude API, Claude Code, and Claude Cowork are still in a partial failure state.Subsequently, Grok also went down. xAI marked Grok Web as "outage" from 21:30 Beijing time, and recovery has not yet been announced.Now OpenAI has also joined the list of failures. At 22:58 Beijing time, the OpenAI status page added an incident for "increased error rates for ChatGPT and Codex," stating that they are investigating the issue.

According to Bloomberg, the European Commission has classified OpenAI's ChatGPT as a "very large online search engine," and listed Reddit and Roblox as "very large online platforms," citing that their monthly active user count in the EU has exceeded 45 million.These platforms are required to enhance their handling of illegal and harmful content under the Digital Services Act (DSA), submit transparency reports, risk mitigation plans, and pay annual fees to the EU, with violations potentially incurring fines of up to 6% of global annual revenue. The DSA also mandates platforms to strengthen protections for minors and advertising transparency, and EU regulatory authorities have initiated over a dozen investigations into various online platforms based on this law.

On August 25, OpenAI announced the addition of proxy browsing capabilities for ChatGPT Work, enabling it to take over tasks on websites that require login and continue working. Users only need to enter their credentials or security codes once at the login interface, and the proxy can continue executing tasks, with sessions potentially remaining logged in for subsequent tasks without the need to log in again. OpenAI stated that this browser supports password managers, and the model cannot see usernames or passwords; this information will not be stored or used for training.However, one-time authentication effectively leaves a foothold for the proxy to maintain continuous access to the account. OpenAI's own models have previously exhibited boundary-crossing behavior, with approximately 1,200 OpenAI proxies (including GPT-5.6 Sol and a pre-release model) breaching the testing environment and invading Hugging Face's production servers to cheat benchmark tests, with about 700 proxies participating in the attack. This feature has been launched on the web and mobile browsers of ChatGPT Work, allowing users to clear sessions by site individually in the cloud browser settings.

On August 26, B.AI announced that the API is fully compatible with the OpenAI Responses API. Developers only need to configure the Base URL using the B.AI API Key to directly call the GPT series models supported by B.AI in the Codex development environment, seamlessly integrating cutting-edge large model capabilities into their daily coding workflow. This means that whether it's code generation, logical understanding, intelligent debugging, architectural refactoring, or team development collaboration, developers can obtain real-time intelligent assistance from the GPT series models in the Codex environment, comprehensively covering various scenario needs from individual development to team collaboration. Experience it now: http://b.ai

According to X user synthwavedd, OpenAI has recently completed the next phase of pre-training, internally codenamed "Bel," which is the successor model to the previous "Doug." This model has a total parameter count exceeding 10T, comparable in scale to GPT-4.5. It is expected that after further reinforcement learning, it will become the foundational model for Astra and GPT-6, and is anticipated to serve as the base model approaching the AGI threshold in the post-GPT-6 era.Reports indicate that this development has caused concern for Anthropic. OpenAI believes that Anthropic currently lacks an effective plan to respond to the public release of Astra; insiders claim that this assessment is largely accurate, primarily due to limited computing power. Anthropic is preparing for the possibility that OpenAI will dominate for the remainder of this year, but it is expected to regain a lead early next year.

Recently, multiple users discovered an image model named luna-lisa-alpha appearing on the LMArena testing platform, suspected to be a new generation image generation model that OpenAI is testing. Although not officially confirmed, there is speculation that it could be an early version of GPT Image 2.5 or the next generation of GPT Image.From the testing feedback, the new model shows significant improvements in character consistency, image realism, text generation capability, and noise control. Some testers compared it with models like GPT Image 2.0, Nano Banana Pro, and Grok Imagine, believing that luna-lisa-alpha can better maintain the appearance, clothing, and style of the same character across multiple generations. However, some testers pointed out that there are still issues with repeated textures in complex texture areas. On the same day, OpenAI announced that the GPT Image 2 API preview supports transparent background generation, which can be used for product images, web design, and other scenarios. The exposure of multiple checkpoint names has also led the community to speculate that OpenAI may be conducting A/B testing. Whether this model will be officially released remains uncertain, and we need to wait for official information from OpenAI.

At the 2026 World Robot Conference, Wang Xingxing, founder and chairman of Yushu Technology, delivered a keynote speech titled "From Exhibits to Products: The Next Decade of the Humanoid Robot Industry." He stated that the company has been continuously promoting the integration of robots into daily life in recent years, including applications in automotive factories, deploying simple tests in its own factory, and the AI team advancing robots into households for work. However, due to overall efficiency and versatility still being insufficient, large-scale promotion has not yet occurred.Wang Xingxing believes that moving towards an embodied intelligent ChatGPT moment ideally requires 2 to 3 years, or even longer, 5 to 10 years. Reaching this moment means that in 80% of unfamiliar scenarios, robots can successfully complete about 80% of tasks through voice or text commands. The most critical challenge currently is the alignment of input and output of AI large models with real robots.

OpenAI announced the expansion of its cybersecurity defense program Daybreak and launched the GPT-5.6-Cyber model specifically for the cybersecurity field, aimed at helping authorized security researchers and defense teams enhance their vulnerability discovery, threat analysis, and security testing capabilities. As attackers increasingly leverage AI to launch faster and larger-scale cyberattacks, defenders need to gain advanced AI capabilities in advance.This Daybreak offers two types of access: Daybreak Blue is aimed at most defense teams, providing general models such as GPT-5.6 Sol for vulnerability discovery, secure code review, malware analysis, incident response, and patch validation; Daybreak Red is aimed at advanced security research, providing GPT-5.6-Cyber for authorized vulnerability research, vulnerability validation, and security testing. GPT-5.6-Cyber is trained on GPT-5.6 Sol and optimized for cybersecurity tasks, including discovering zero-day vulnerabilities and analyzing exploit chains. It has been used in actual vulnerability research and has identified high-risk vulnerabilities in software, including the Chrome V8 JavaScript engine. Additionally, OpenAI stated that GPT-5.6-Cyber has also helped discover high-risk vulnerabilities in various domains, including privilege escalation vulnerabilities in mobile operating systems, remote code execution vulnerabilities in databases, and hundreds of privilege escalation vulnerabilities in operating system kernels.OpenAI also emphasized that Daybreak Red will only be accessible to approved individuals and organizations, controlling access through measures such as authentication, account security, monitoring, usage restrictions, and legal disclaimers. The company stated that it will continue to strengthen security monitoring, access management, and model security testing to reduce the risk of advanced cybersecurity models being abused.