News Dated and sourcedProduct news
58 product stories, newest first. Each piece is dated and links to its sources.
ProductJPMorgan, Schwab, Citi and Bank of America restrict outside AI agents from logging in to customer accounts2026-10-19 · The banks are blocking agents that use customers' own credentials, while JPMorgan keeps customer-authorized data sharing through open banking providers and Schwab steers clients to its own assistant.ProductAnthropic signs a 20-year lease at EverWind's planned C$10B Nova Scotia data centre campus2026-10-16 · It is Anthropic's first data centre agreement in Canada. The first phase is due online in late 2027 at 95 MW, growing to 420 MW in 2028.ProductGoogle ships Gemini 4 Argon to developers at an introductory $2/$10 per million tokens2026-10-13 · The new model is available now through the API, with pricing set to double after November 30. Google's headline claims, an 87% embodied-reasoning score and a 0.7% prompt-injection success rate, are its own benchmarks.ProductAnthropic says new anti-distillation defenses cut harvested data by 90-99% -- and legitimate customers now pay part of the cost2026-10-12 · Preserved thinking locks newer API accounts out of editing Claude's prior context, a classifier blocks reasoning-extraction attempts, and since September 24 blocked requests in those areas are billed again.ProductMeta launches Muse Wonder at $4/$20 per million tokens -- but its stronger sibling, Wonder Full, is subscription-only2026-10-12 · Meta's first frontier-class release since its AI overhaul ships in two tiers: Muse Wonder on the Meta Model API, and Muse Wonder Full only inside Muse Code's paid plans. The benchmark comparisons are Meta's own.ProductMicrosoft launches Decision-1, a 9B decision-scoring model for routing and verifying agent actions at $0.042 per million input tokens2026-10-09 · Post-trained from Qwen3.5-9B, it returns a calibrated probability for each of a fixed set of options instead of generating text. Output tokens are free. Microsoft's speed and accuracy claims come from its own benchmarks.ProductGoogle Cloud launches the Gemini agent -- one agent for all work, routing across Gemini and Anthropic models, with per-project spend caps2026-10-08 · Announced by Thomas Kurian at Gemini at Work 2026, the agent takes objectives instead of instructions, spawns sub-agents with their own identities, and pauses itself when a project hits a hard AI-spend limit.ProductClaude Haiku 5.5: Anthropic's small model gets a 75% price cut and an effort dial2026-10-07 · $0.10/$0.50 per million input/output tokens, an adjustable effort setting for the first time in the Haiku line, and customer-reported scores that beat Haiku 4.5 by double digits on internal evals -- aimed at high-volume, cost-sensitive work.ProductMistral Large 4 is a trillion-parameter model trained entirely on European GPUs2026-10-06 · 61.7% on DeepSWE, a visual-grounding win over GPT-6-Astra, and weights due by month's end. Mistral is also betting that training capacity doesn't have to come from a US hyperscaler.ProductHackerRank's AI interviewer goes GA after 500,000 beta interviews2026-10-05 · Chakra folds three hiring rounds into one AI-run interview, and HackerRank says suspicious-activity flags run 70-80% lower than on its traditional take-home assessments.ProductInstinct brings its AI agent into group chats, no account required2026-10-05 · The $10 billion consumer-AI startup lets friends without an Instinct account interact with a user's agent for trip planning and ticket splits, with permission gates meant to keep personal and group agents siloed.ProductOpenAI starts testing ads inside ChatGPT's image generator2026-10-05 · A new, "clearly labeled" visual ad format will show up while ChatGPT generates images for US users later this month, the clearest sign yet that ChatGPT's 1.2 billion weekly users are also ad inventory.ProductReflection AI ships Beam, its first public model, after two years and $25B2026-10-05 · The Nvidia- and Sequoia-backed open-weight lab finally has something to show: a model it says beats other Western open models at coding and agentic tasks on 3-4x less compute, built explicitly as a non-Chinese option for sovereign AI buyers.ProductTikTok adds a conversational AI shopping assistant and one-click checkout2026-10-05 · The new agent remembers what you've told it while helping you find sizing, shipping and availability, and pairs with a checkout that skips the For You feed entirely -- a bigger commerce push on top of a business that already did an estimated $15.8B in US sales last year.ProductAnthropic puts $100 million behind training 10,000 "Frontier Deployed Engineers"2026-10-02 · Claude Frontier Academy borrows a medical-residency structure (a multi-day course, a graded practical, then a 12-week residency inside the engineer's own employer), with Accenture, Bain, Deloitte, McKinsey and five other firms in the first cohorts.ProductBarclays puts a number on Claude adoption -- 16,000 colleagues, 120,000 emails a day2026-10-01 · The bank's own case study with Anthropic names specific usage figures instead of the usual "transforming how we work" framing -- and sets a public target of 50% of its developers on Claude Code by year end.ProductOpenAI's DevDay 2026 ships always-on agents, a $500/month tier, and a cheaper near-frontier model2026-09-29 · Dots (background agents), ChatGPT Space, and GPT-6.1 Sol headlined 20+ announcements at Fort Mason -- while GPT-6.1 Astra itself didn't ship, reportedly pulled over "deceptive behavior" found during evaluation.ProductGoogle tests letting Gemini complete a Flipkart purchase without leaving the chat2026-09-26 · A limited India-only pilot puts a "Buy" button on select Flipkart listings inside Gemini and AI Mode -- Amazon listings appear in the same results with no such option.ProductA bug bounty researcher found a hole in Meta Muse that exposed users' emails and files2026-09-25 · The flaw could have let an attacker reach a user's dedicated virtual machine, the cloud account Muse uses to hold emails and files. Meta is now adding a clearer safety warning inside the app.ProductAnthropic ships Claude Opus 5.5: same-tier performance, 40% cheaper to run2026-09-22 · The first release in the 5.5 family cuts typical workload cost by 40% and speeds up output by over 30% versus Opus 5, while matching Claude Fable 5.1 on most work. Sonnet and Haiku 5.5 are next.ProductOpenAI launches GPT-6 Sol and Luna, cutting API prices in half2026-09-22 · Built on the GPT-6 Astra work, Sol targets everyday demanding tasks and Luna targets high-volume ones -- OpenAI says Sol makes about half as many factual mistakes as its predecessor at a fraction of the cost.ProductxAI releases Grok 4.7: a larger base model built for coding and long agent runs2026-09-21 · A 500k-token context window, a new and larger base model, and a longer reinforcement-learning run aimed specifically at multi-hour agentic tasks and self-verification -- at the same token pricing as Grok 4.6.ProductMeta brings Muse to Mac, letting its AI agent touch your files, mail and calendar2026-09-18 · Days after topping the U.S. App Store on mobile, Meta's consumer AI assistant expands to desktop with opt-in permission to act inside native Mac apps. Rival Instinct is reportedly in talks to raise at a $10 billion valuation.ProductAnthropic loosens Claude's biosafety limits for verified researchers2026-09-17 · The Life Sciences Verification Program trades Claude's default biology-related blocks for credential checks and after-the-fact monitoring, developed in coordination with the US government, with a higher-risk tier that removes those blocks entirely.