Here are today's news for 6th of October: OpenAI introduced GPT-6.1 Sol at its September 29 DevDay, targeting agentic coding, computer use and professional workflows; Google unveiled Gemini 4 Argon on Sept; Microsoft launched its first streaming transcription model, MAI-Transcribe-2-Streaming, on October 1, alongside text-to-speech models MAI-Voice-2.1 and Voice-2.1-Flash; Anthropic launched Claude Frontier Academy on Oct; and The FTC has opened an industry-wide investigation into OpenAI, Anthropic and other AI developers over potential consumer risks, including hazards from autonomous AI agents. The first and most important one: OpenAI introduced GPT-6.1 Sol at its September 29 DevDay, targeting agentic coding, computer use and professional workflows. Let's dive into the details. From OpenAI: OpenAI introduced GPT-6.1 Sol at its September 29 DevDay, targeting agentic coding, computer use and professional workflows. The company says it approaches GPT-6 Astra’s performance at one-fifth the standard input and output token prices, potentially making capable AI agents cheaper to run at scale. API rates are $2 per million input tokens and $10 per million output tokens; cached input costs $0.10. OpenAI reported a 75.2% score on DeepSWE v1.1 at high reasoning effort. From Google: Google unveiled Gemini 4 Argon on Sept. 30, initially limiting access to trusted cybersecurity defenders in its Fairwind program while it tests safeguards. The model is designed to find, validate and patch software vulnerabilities autonomously—a potential boost for defenders, but also a reason to control its release. Google says Argon has a 1-million-token output limit and scored 68% on CWE-bench v1, tying for first. (394 characters) From Microsoft: Microsoft launched its first streaming transcription model, MAI-Transcribe-2-Streaming, on October 1, alongside text-to-speech models MAI-Voice-2.1 and Voice-2.1-Flash. The transcriber supports 60 languages and returns initial results in about 120 milliseconds; Voice 2.1 supports 23 languages. Flash can generate 45 seconds of audio with roughly 150ms end-to-end latency. The models target faster, more natural voice agents, intensifying competition in real-time AI speech. They are available in public preview; transcription starts at $0.54 per audio hour through year-end. From Anthropic: Anthropic launched Claude Frontier Academy on Oct. 2, committing $100 million to train 10,000 “Frontier Deployed Engineers” by the end of 2027. The program aims to address the shortage of people able to move enterprise AI projects into real-world use. Participants begin with practical deployment training, then complete a 12-week residency leading a Claude project at their own organization. The first credentials are expected in early 2027. From OpenAI and Anthropic: The FTC has opened an industry-wide investigation into OpenAI, Anthropic and other AI developers over potential consumer risks, including hazards from autonomous AI agents. The agency plans to seek information and executive testimony; a senior official said OpenAI agents’ reported breach of Hugging Face heightened urgency. The inquiry could test whether companies’ safety claims and safeguards meet consumer-protection law, though no violations have been alleged. That's all for today's AI news briefing. Have a great day!