Here are today's news for 27th of October about OpenAI, Google DeepMind, Meta and Anthropic. Let's dive into the details. OpenAI released GPT, 6.1 Terra on October 26, 2026, positioning it as a fast, efficient model for visual reasoning, computer use and writing. It is rolling out in ChatGPT Work and Codex for Plus, Pro, Business, Enterprise and Edu users; developers can call it through the API as `gpt, 6.1, terra`. API rates are $1 per million input tokens, $0.05 per million cached input tokens and $5 per million output tokens. OpenAI says Terra scores 15 percentage points above GPT, 5.6 Terra on Agents’ Last Exam at medium reasoning effort, at about one, third the cost, and matches GPT, 6 Astra’s performance at one, tenth the price per token. It also completes tasks more than three times faster than Astra and GPT, 6.1 Sol. For command, line coding, it solved 50.3% of Terminal, Bench 4.0 tasks at maximum effort, versus 21.5% for GPT, 5.6 Terra. The launch extends OpenAI’s push into workplace agents: GPT, 6 Astra arrived in September with computer, use and multistep, work capabilities, followed by GPT, 6.1 Sol later that month. On October 9, 2026, Hurricane Isaias came ashore near Destin, Florida, as a Category 2 storm with 105, mph winds. Google DeepMind’s WeatherNext Cyclones had forecast a Florida Panhandle landfall while conventional guidance initially pointed toward Louisiana or Mississippi. In a comparison of forecasts initialized October 6, the AI placed landfall between Petit Bois Island, Mississippi, and Seaside, Florida, closer to Destin than the European ensemble’s Louisiana projection; it also predicted landfall timing within hours. Its intensity estimate, however, was too low: it forecast 55, knot winds at landfall, versus roughly 95 knots observed. The National Hurricane Center’s blended intensity forecast had 20, 40% more skill than DeepMind’s, underscoring that one storm does not establish overall superiority. The stakes are practical: better track guidance can help coastal residents and emergency managers target preparations, but forecasters still need to weigh AI against physics, based models. Meta introduced Muse Wonder and its larger sibling, Muse Wonder Full, on October 12, its first frontier, model release since its AI reorganization in August 2025. Meta says Wonder Full led rival models on two long, horizon coding tests: it scored 79.0 on DeepSWE 1.1, versus 74.2 for Claude Opus 5.5 and 74.1 for GPT, 6 Astra, and 57.5 on SWE, Marathon, versus 46.3 and 42.5. Results are not uniformly dominant: on Terminal, Bench 4.0, Wonder Full scored 57.1, below Opus 5.5’s 59.6 and Astra’s 59.1. The models accept text, images, video and audio, support a one, million, token context window, and return text; Meta highlights Active Memory and Active Focus for recalling context and persisting through difficult tasks. Wonder is available via Meta’s API at $4 per million input tokens and $20 per million output tokens; Wonder Full is limited to Muse Code subscribers. The launch builds on Muse Spark, which Meta positioned as a faster, lighter model, and advances its effort to compete with OpenAI and Anthropic. Anthropic filed its U.S. IPO registration on Oct. 26, 2026, revealing first, half revenue of $16.29 billion, up from $1.16 billion a year earlier, and a net loss of $80.7 billion. The headline loss is largely noncash: $72.4 billion came from revaluing convertible notes and related instruments as Anthropic’s valuation rose. Its operating loss was $693 million, while adjusted operating income reached $1.62 billion; second, quarter revenue was $11.6 billion. The figures expose the gap between surging demand for Claude and the cost and financing complexity of building frontier AI. Infrastructure spending approached $9 billion in the first half, compared with $7.3 billion for all of 2025, as the company competes with OpenAI and Google for customers and computing capacity. Anthropic’s Claude models and services have driven rapid growth, with revenue 13.5 times the year, earlier half. The filing gives prospective investors a detailed view of growth, costs and risks ahead of a potential listing; no share count or price range has been set. On October 26, Meta launched roughly $35 billion in underwritten stock offerings: $22.5 billion of Class A common shares and about $12.5 billion in depositary shares tied to mandatory convertible preferred stock. It also plans a separate, up, to, $60 billion at, the, market share, sale program after its non, voting Class C shares begin trading, expected in the first quarter of 2027, potentially bringing the equity raise to $95 billion. Nvidia, AMD and Broadcom have expressed nonbinding interest in buying up to $5 billion of the Class A shares combined. The financing underscores the scale of the infrastructure race: Meta forecasts 2026 capital expenditure of $135 billion to $145 billion as it expands data centers and computing capacity for AI products, including its apps and Muse agent, and its push for “personal superintelligence.” That spending puts Meta alongside rivals such as OpenAI and Anthropic in competing for chips, power and data, center capacity. That's all for today's AI news briefing. Have a great day!