Latest AI News: Fable 5.2 & Wallet (Nov 15)

Latest AI news: Claude Fable 5.2, Gemini 4 Hydrogen, ChatGPT Wallet for agent shopping, and Nvidia CEO exits. By AISA's AI agents.

By AISA Team··9 min read
AI newsweeklylatest AI news

Claude Fable 5.2 took the top spot on the Artificial Analysis Intelligence Index this week, Google shipped two Gemini 4 models in three days, and OpenAI gave its AI agents a wallet to shop with.

This week's AI news covers one of the densest release windows in recent memory: six model launches, a payment system for AI agents, a full office suite from Anthropic, and a CEO departure that rattled the industry's largest chipmaker. Here is what happened and what it means for practitioners.

Claude Fable 5.2 Takes the AI Benchmark Crown

Anthropic released Claude Fable 5.2 on November 9, its most capable publicly available model. It scores 65 on the Artificial Analysis Intelligence Index at max effort, up from 50 for Fable 5.1 — a 15-point jump in one generation.

The model keeps the 1M-token context window and 128K max output of its predecessor. API pricing dropped 20% to $8 per million input tokens and $40 per million output tokens. Output is up to 30% faster than Fable 5.1, and a new fast mode delivers up to 2.5x speed at double the price ($16/$80). The API identifier is claude-fable-5-2, with a knowledge cutoff of June 2026 and adaptive thinking always on.

Alongside Fable, Anthropic released Claude Mythos 5.2 — the same underlying model with fewer safeguards, available only to verified cybersecurity and life-sciences organizations. Understanding the difference between these tiers matters for model selection in enterprise settings.

Google Ships Gemini 4 Carbon and Hydrogen in Three Days

Google made Gemini 4 Carbon generally available on November 10, followed by Gemini 4 Hydrogen on November 12. The two models serve very different use cases.

Gemini 4 Carbon

Carbon is Google's strongest reasoning model, designed for cybersecurity defense and long-running agentic coding. Its headline feature is a unified 1,048,576-token window shared by input, thinking, and output — meaning a response can run past 64K tokens up to the full window minus prompt size. It scores 60 on the Artificial Analysis Intelligence Index, currently ranking #1.

Introductory pricing runs at $2 per million input tokens and $10 per million output tokens until November 30, after which it doubles to $4/$20. The model accepts text, image, audio, and video input. API identifier: gemini-4-carbon.

Gemini 4 Hydrogen

Hydrogen is Google's fast everyday model — the Gemini 4 replacement for 3.8 Flash. Google describes it as their "daily driver for agentic workflows." It completes tasks 2.5x faster than Flash at its default reasoning setting and scores 51–54 on the Intelligence Index depending on thinking level (low, medium, high, or the new xhigh tier).

Priced at $1.50/$7.50 per million tokens, Hydrogen is available on free tiers via automatic model selection, with named access starting at Google AI Plus ($4.99/month). This model matters for anyone building agentic workflows that need speed without frontier-level cost.

Meta Muse Ember Targets AI Agent Workloads

Meta released Muse Ember on November 9, its smallest and most efficient model. Priced at $0.10 per million input tokens and $0.50 per million output tokens, it matches Claude Haiku 5.5 and GPT-6 Luna on price while scoring 36 on the Intelligence Index — level with Haiku at roughly one-third the cost per task ($0.14 vs $0.41).

Ember takes text, images, audio, and video input with a 1M-token context window and up to 512K output tokens. Meta positions it for the repetitive work agents do all day: tool calls, coding subtasks, routing, extraction, and document parsing. The model is designed to pair with Meta's larger Muse Wonder model in agent harnesses — Wonder plans, Ember handles volume. Model ID: muse-ember-1.0. It is hosted-only; no open weights have been released.

ChatGPT Wallet Lets AI Agents Shop and Pay

OpenAI introduced ChatGPT Wallet on November 13, a payment feature that lets its "dots" agents buy things on behalf of users. This is the first major AI agent payment system from a frontier lab.

The flow: users save an eligible card once, then their dot agent finds products, compares options, and presents the merchant, items, total, and card for approval. Each purchase requires biometric confirmation (Face ID or Touch ID). The agent never sees actual card details — the payment network provides a tokenized credential. OpenAI built it in partnership with American Express, Mastercard, and Visa.

The rollout is limited to US ChatGPT Pro subscribers through the dots experience. OpenAI plans to add spending limits, category controls, and pre-approved purchases in future updates. This development has direct implications for anyone studying human-in-the-loop patterns and agent trust boundaries.

Anthropic Launches an AI Office Suite

On the same day as Gemini 4 Carbon's GA release, Anthropic launched a suite of workplace tools that position Claude as a direct competitor to Google Workspace and Microsoft 365.

Claude Sheets

A collaborative spreadsheet tool in beta for paid plans. Claude builds multi-tab files with formulas and dropdown menus from prompts, highlights every cell edit it makes, and supports shared editing with Excel/CSV export.

Claude Meetings

A Google Meet agent (beta for Team and Enterprise plans) that joins calls as a visible participant, takes notes, tracks decisions, and produces action items before the call ends. Whiteboards can build alongside the conversation in real-time. Claude auto-exits when uninvited external guests join — a privacy gate absent from competing meeting tools.

Claude Whiteboards and Data

Whiteboards (beta) offer a collaborative visual canvas for planning sessions. Claude Data (formerly Dashboards) went generally available, providing analytics capabilities within the Claude workspace.

AISA

Curious about your AI Fluency?

AISA helps you measure, prove and improve your AI skills — free report in a 20–40 minute chat.

AI Security: Grok Build Exposes 35,000+ Apps

Security firm Claroc found that 35,578 of 37,571 Grok-built apps (94.7%) were accessible without sign-in, exposing internal data from SpaceXSI, X, Fortune 500 companies, and a defense contractor. The vulnerability stems from Grok Build's link-sharing mechanism defaulting to "anyone with the link" when publishing.

The exposed apps included 24 built by SpaceXSI and X staff, containing sales territory maps, onboarding plans, and internal system addresses. Claroc reported the issue on November 8; SpaceXSI began removing links after inquiries.

This incident is a practical case study for prompt security and reinforces why stakes-based review processes matter when deploying AI-generated applications. Teams using any AI app builder should audit their default sharing settings immediately.

Jensen Huang Resigns as Nvidia CEO

Jensen Huang resigned as Nvidia's President, CEO, and board member on November 12, effective immediately — ending a 33-year tenure at the company he co-founded. The resignation followed reports of extramarital affairs, one involving a woman suspected by US intelligence of links to Chinese espionage.

Colette Kress, Nvidia's CFO, was appointed interim CEO. The board has begun searching for a permanent successor and reportedly approached Arm CEO Rene Haas. Nvidia reports quarterly results on November 17, its first without Huang. Three House committees have summoned Huang for a closed-door interview on December 1.

METR Finds AI R&D Acceleration at Anthropic

METR, the independent evaluation nonprofit, published an 11-week assessment on November 5 concluding that Anthropic's model capabilities are improving at roughly 1.9x the historical rate (March 2024 to October 2025 baseline). METR attributes approximately 80% of the acceleration to Anthropic's internal use of its own AI systems in research and development.

The gains came primarily through AI-assisted work rather than fully autonomous processes. METR ruled out training-compute scaling as the explanation. The assessment also revealed that Anthropic's internal "Model 4" is the first to clear 50% on every METR benchmark. This is early quantified evidence of AI-assisted recursive improvement at a frontier lab, with significant implications for how quickly model benchmarking baselines may shift.

Other Notable Developments

  • Adobe MAX (Nov 10): Adobe expanded its Firefly Creative Agent with Plan Mode (shows its plan before executing), Model Selector (picks from Adobe, Google, OpenAI, and Kling models), Firefly Graph (reusable node-based workflows, beta), and a local MCP for Creative Cloud apps.
  • OpenAI dissolved its robotics team, moving researchers into the main research group to focus on operating robots with general-purpose models.
  • CrowdStrike acquired two AI agent security startups for $495M.
  • IBM and Microsoft launched Imagine Enterprise, a platform combining prebuilt AI agents and cloud infrastructure.
  • Anthropic's IPO is expected imminently, with the company valued at $965 billion.

What This Means for Your AI Skills

This week's releases shift the practical landscape in several ways. The pricing drops on Fable 5.2 and the arrival of budget models like Muse Ember mean that model comparison is no longer just about capability — cost-per-task at different reasoning levels is now a core skill. ChatGPT Wallet introduces a new category of agent interaction that demands understanding of trust boundaries and human oversight patterns. And the Grok Build security incident is a reminder that AI fluency includes knowing what can go wrong.

AISA's conversational AI reads this snapshot every week to stay current on the AI landscape, so when you take the assessment, your conversation reflects the latest developments — not a frozen training set. Whether you are exploring these new models or building agent workflows, the AI skills rubric maps these capabilities to measurable proficiency levels.

Take the free AI skills assessment to see where you stand.


Related reading: Top 10 AI Skills Certifications in 2026 — Which credentials actually matter this year.

Related reading: AI Fluency: The New Digital Literacy — Why understanding AI tools is becoming a baseline professional skill.

Related reading: How Good Are Most People at AI? — Benchmark data on where professionals actually score.

Frequently Asked Questions

What are the biggest AI developments this week?

The biggest developments for the week of November 9–15, 2026 are Claude Fable 5.2 taking the #1 spot on the Artificial Analysis Intelligence Index, Google releasing Gemini 4 Carbon and Hydrogen within three days, and OpenAI launching ChatGPT Wallet for AI agent payments. Jensen Huang's resignation from Nvidia also dominated industry news.

Which new AI models launched this week?

Four notable models launched: Claude Fable 5.2 (and Mythos 5.2) from Anthropic on November 9, Meta Muse Ember on November 9, Gemini 4 Carbon GA on November 10, and Gemini 4 Hydrogen on November 12. All four offer 1M-token context windows at varying price points from $0.10 to $8 per million input tokens.

How do this week's AI changes affect professionals?

Practitioners now have more model choices across the cost spectrum, from Muse Ember at $0.10/MTok for agent subtasks to Fable 5.2 at $8/MTok for complex reasoning. ChatGPT Wallet introduces agent-driven purchasing that professionals in e-commerce and product management should understand. The Grok Build security incident highlights the importance of auditing default settings in AI development tools.

Ozan Dagdeviren

Ozan Dagdeviren

Founder of AISA — the AI skills assessment platform used by professionals worldwide to measure, certify, and develop their AI fluency. More about AISA

AISA

Curious about your AI Fluency?

AISA helps you measure, prove and improve your AI skills — free report in a 20–40 minute chat.

The Science Behind AISA

Metropolitan PoliceHarvard UniversityCrowdboticsE.S.E.

In 2026, Anthropic published the AI Fluency Index — the largest empirical study of AI fluency to date, analysing nearly 10,000 conversations. AISA covers 93% of the behaviours Anthropic identified as markers of AI fluency and goes even deeper with 4 additional criteria. The U.S. Department of Labor's AI Literacy Framework (TEN 07-25) defines what every worker needs to know about AI — AISA covers 100% of its 25 sub-competencies.Read our analysis: Anthropic's AI Fluency Study & AISA · DOL AI Literacy Framework & AISA

AISA's framework is developed by a team with deep roots in tech, behavioural science, and AI product leadership — the rubric is informed by backgrounds spanning the Metropolitan Police, Harvard, Crowdbotics (Silicon Valley), and the European School of Economics.