🤖 GPT-5.5 Instant ships as ChatGPT's new default 🤖
OpenAI's rollout cuts hallucinations in half but raises API pricing considerably.

OpenAI has released GPT-5.5 Instant as the new default model powering ChatGPT, delivering a 52.5% reduction in hallucinations across high-stakes domains like medicine, law, and finance. The model also brings stronger personalization based on user context and more concise responses. On the pricing side, the update comes with a 2x nominal price increase over GPT-5.4, though fewer completion tokens on longer prompts partially offset the jump — real cost increases range from 49% to 92% depending on use case. This sets the pricing bar for consumer AI defaults. The release reflects OpenAI's accelerating model iteration cadence, reinforcing its grip on the consumer AI market while pushing enterprise users to weigh accuracy gains against rising costs.

🏦 Anthropic ships 10 AI agents for finance 🏦
Anthropic's integration of ready-to-run agents targets banks and insurers directly.

Anthropic has launched 10 AI agent templates purpose-built for financial services and insurance, covering tasks from pitchbook creation and KYC file screening to earnings reviews and month-end book closings. Each agent ships with domain-specific skills, data source connectors, and add-on Claude models for sub-tasks. They run as plugins within Claude Cowork, Claude Code on desktop, or as Managed Agents on the Claude platform. Alongside, Anthropic announced a Claude add-in for Microsoft 365 and new data connectors from partners including Dun & Bradstreet, Verisk, and IBISWorld. This is Anthropic's most aggressive push into vertical enterprise tooling — a direct signal of competitive pressure in the race to own regulated industry workflows.

⚡ Subquadratic opens 12M-token context window model ⚡
A startup's speed claims could reshape assumptions about long-context infrastructure costs.

AI startup Subquadratic has unveiled SubQ, billed as the world's first fully sub-quadratic frontier model, featuring a 12-million-token context window that outperforms GPT-5.5 on retrieval benchmarks. The sparse-attention architecture reportedly requires 1,000x less compute than competing models, directly tackling the quadratic cost scaling that has bottlenecked long-context processing. The company has already announced plans for a 50-million-token model. The launch video has surpassed 7 million views, signaling strong developer interest. If efficiency claims hold at scale, this could slash infrastructure costs for long-context applications — a meaningful challenge to incumbents betting heavily on brute-force compute.

INTERESTING TO KNOW

🚀 Gemma 4 gains 3x speed via multi-token drafters 🚀

Google's rollout of Multi-Token Prediction drafters for the Gemma 4 model family delivers up to a 3x inference speedup with zero degradation in output quality or reasoning. The approach uses speculative decoding — predicting several future tokens simultaneously while utilizing idle compute cycles. This directly boosts competitive positioning for open-weight model deployment speed, giving developers building real-time applications a compelling reason to choose Gemma 4 over proprietary alternatives.

💰 Anthropic commits $200B to Google Cloud 💰

Anthropic has locked in a $200 billion pricing commitment to Google Cloud infrastructure over five years — a deal now representing over 40% of Google's revenue backlog. Google plans to invest up to $40 billion in Anthropic in return. The massive spend reflects surging compute demands from Claude model usage, which has led to capacity constraints and user-facing usage caps. This deepens cloud provider dependency for frontier AI labs, and significantly strengthens Google Cloud's position against AWS and Azure.