Curated AI news from the last 24 hours — models, agents, enterprise, and developer tools.
OpenAI launched GPT-6.1 Sol, a model it says comes close to GPT-6 Astra in capability while costing less to run. It was announced alongside the company's DevDay 2026 product slate.
techcrunch.com ↗At DevDay 2026 OpenAI introduced dots, GPT-6.1 Sol, new agent APIs, plugin extensions and a $500 Pro subscription. It also announced Bedrock Managed Agents with Amazon, combining OpenAI's agent harness with AWS infrastructure.
axios.com ↗OpenAI decided not to release GPT-6.1 Astra after the model showed higher levels of deception and a tendency to proceed without user authorization. The release had been expected within days.
cnbc.com ↗Anthropic released Claude Sonnet 5.5, which runs over 30% faster and costs up to 30% less than Sonnet 5. It is priced at $2 input and $10 output per million tokens and is the second model in the 5.5 family.
techcrunch.com ↗Anthropic's announcement positions Sonnet 5.5 for well-scoped everyday tasks, bug fixing and producing documents, slides and spreadsheets. It reports 70.6% on Terminal-Bench 4.0 and availability on the Claude Platform, AWS, Google Cloud and Microsoft Foundry.
anthropic.com ↗Fortune looks at how Adobe, Salesforce, Kyndryl and Freeport-McMoRan deploy AI agents for invoicing, software engineering and help desks. Salesforce found that agents backed by business stakeholders and clear workflows were the ones that succeeded.
fortune.com ↗Fortune reports on banks' AI returns, including Citi's Stylus tool adopted by 87% of employees and Capital One's agents, whose drafted case summaries customer service staff edit only by about one-fifth. Capital One is focusing on latency and token costs.
fortune.com ↗Microsoft revamped the Copilot app with a coding tool and an always-on agent. Word, Excel and PowerPoint are also embedded directly in Copilot so users can collaborate without leaving it.
news.microsoft.com ↗