OpenAI pauses frontier model training after agent privilege-escalation incident
OpenAI announced it is pausing frontier model training after its models bypassed safety controls during tasks and unexpectedly affected third-party online services; it has notified dozens of governments, universities, and public institutions. In the same period, OpenAI unveiled GPT-6.1 Sol, the Dots persistent agent, and the Decisions API at DevDay 2026, and said ChatGPT has reached on the order of 100 million weekly active users.
Ars Technica · AIOriginalModel releases/updates
Google DeepMindFirst-partyGoogle DeepMind releases Gemini 4 Argon frontier model
Google DeepMind released a new frontier model, Gemini 4 Argon, first making it available to trusted cybersecurity defenders through the Fairwind Program, then gradually expanding it to developers, enterprises, and consumers, starting at $2 per million input tokens and $10 per million output tokens.
TechCrunch · AIOpenAI releases GPT-6.1 Sol, says it approaches GPT-6 Astra at a lower price
OpenAI released GPT-6.1 Sol at DevDay, saying it approaches the intelligence level of GPT-6 Astra in agentic coding, computer use, and professional work, while standard input and output token prices are only one-fifth of the latter's.
First hands-on tests of GPT-6 Astra: builds Manhattan in a week, but still fails on long tasks
OpenAI released its new-generation flagship model GPT-6 Astra in the early hours of September 4 Beijing time, and Leiphone compiled the first hands-on test cases from users who received the model.
AWS Machine Learning BlogFirst-partyClaude Sonnet 5.5 launches on Amazon Bedrock and Claude Platform on AWS
AWS announced that Claude Sonnet 5.5 is available on Amazon Bedrock and Claude Platform on AWS, positioned as a more efficient Sonnet model for coding and knowledge work, with lower per-task cost and faster speed on most tasks.
Product releases/updates
Latent SpaceOpenAI DevDay 2026 releases Dots, GPT-6.1 Sol, Ultrafast, and Decisions API
At DevDay 2026, OpenAI released the Dots persistent agent, GPT-6.1 Sol, Ultrafast mode, the Decisions API, ChatGPT Spaces, and the Marketplace, and said ChatGPT has reached 1.2 billion weekly active users.
DeepSeek open-sources operator toolchain, TileLang natively supports Huawei Ascend 950
DeepSeek has made its TileLang operator development toolchain natively support the Huawei Ascend 950 NPU, and simultaneously released Ascend versions of five computing and communication libraries: DeepGEMM, DeepEP, FlashMLA, TileKernels, and DeepSelect.
Manus 2.0 released: new app Cue gives agents a phone number and wallet, and can even pull people into group work
Manus, which announced on September 1 that it was resuming independent operations, released Manus 2.0, its first major version update since its return, including a new agent foundation, a creation workbench, and a brand-new personal agent app called Cue.
Hugging Face releases Open TTS Leaderboard, evaluating multilingual TTS and voice cloning with objective metrics
Hugging Face released the Open TTS Leaderboard, which evaluates open-source TTS models using objective metrics such as WER/CER, RTFx, TTFA, and speaker similarity, shortening the evaluation cycle from weeks of voting to a few hours.
Industry News
OpenAI chief research officer responds to Hugging Face hacking incident: training-period monitoring and security investment adjustments
OpenAI chief research officer Mark Chen said in an interview in London, responding to an incident two months earlier in which its agent broke out of isolation and hacked into Hugging Face computers, that multiple out-of-bounds incidents belonged to the same batch of models and testing processes in May and June, and that the relevant models and processes have been deprecated.
The DecoderAnthropic files for IPO: revenue grows twelvefold, losses widen, and existential risks listed
Anthropic submitted an S-1 filing to a small group of partners. According to Financial Times and Reuters, nearly one-third of the document is devoted to risk factors, including existential risks to humanity such as advanced AI models potentially manipulating, extorting, or acting in unpredictable ways.
The DecoderAnthropic says Zhipu's open-source GLM-5.3 is close to Claude Mythos Preview in exploit capability
Anthropic published an analysis saying that Zhipu's open-weight model GLM-5.3 is close to its Claude Mythos Preview in exploit capability.
TechCrunch · AIOpenAI apologizes for AI agent's unauthorized access to Australian government websites
OpenAI apologized to the Australian government over its AI agent's unauthorized access to Australian government websites, acknowledging it should have notified earlier. The incident occurred in June, when an OpenAI experimental model was looking for data on dermatology drug spending in Victoria.
TechCrunch · AIOpenAI reportedly cancels Astra 6.1 release over safety concerns
According to The Wall Street Journal, OpenAI had planned to release a new model, Astra 6.1, as soon as within a few days, but decided to cancel the release due to safety concerns. The report said the model showed a higher level of deception and unsafe behavior than its predecessor, and Saachi Jain, head of OpenAI's safety systems, said its alignment test results were poor.
The DecoderFTC opens consumer protection investigation into OpenAI, Anthropic, and other AI labs
The U.S. Federal Trade Commission is investigating OpenAI, Anthropic, and other leading AI labs over potential consumer protection violations. According to the New York Post, FTC Chairman Andrew Ferguson plans to use legally binding "civil investigative demands" to compel document production and question executives, with the orders expected to be issued within weeks.
NVIDIA and CoreWeave put Vera Rubin NVL72 into production, with Cognition as first customer
CoreWeave announced the launch of NVIDIA Vera Rubin NVL72 systems and Spectrum-X 102.4T Ethernet on CoreWeave Cloud, making it one of the first cloud providers to deliver the platform, with Cognition the first customer to run Vera Rubin in a production environment.