Skip to content

Note ai news

All posts
Nov 14, 20267 min readai news · artificial intelligence · anthropic · claude fable · gemini 4 hydrogen · openai · chatgptRead in:ENID

This Week in AI: Anthropic's Trillion-Dollar IPO, Claude Fable 5.2, and Gemini 4 Hydrogen in Search

Anthropic launches Claude Fable 5.2 and kicks off a trillion-dollar IPO roadshow, Google puts Gemini 4 Hydrogen into Search, and OpenAI turns ChatGPT into a business directory.

Close-up of a white robot against a dark purple gradient background, symbolizing AI technology.

This week, Anthropic dominated the AI headlines twice over. The company shipped its new flagship model, Claude Fable 5.2, and opened the roadshow for what could become the largest IPO in history. Google answered with Gemini 4 Hydrogen, a fast model now woven into Search for every user. And OpenAI turned ChatGPT into a place where businesses can be found, claimed, and talked to. Here is what happened, with every claim verified against primary sources.

Key Takeaways

Anthropic launched Claude Fable 5.2 and Mythos 5.2 on November 9, priced 20% lower than Fable 5.1 at $8 per million input tokens and $40 per million output tokens, with vendor-reported FrontierCode 2.0 agentic coding of 49.1%, up from 32.4%. A team of about 1,000 Fable 5.2 agents formalized Fermat's Last Theorem in Lean in 30 hours. (Anthropic)

Anthropic kicked off its IPO roadshow on November 10, seeking to raise up to $64.4 billion at a valuation of about $1 trillion, listing on Nasdaq under the ticker ANT. By the end of Thursday, the book was 6x oversubscribed. (Reuters)

Google launched Gemini 4 Hydrogen on November 12, priced at a flat $1.50 per million input tokens and $7.50 per million output tokens, and rolled it into Google Search globally. It finishes tasks 2.5x faster than Gemini 3.8 Flash and more than triples its Terminal-Bench 4.1 score, 57% versus 18.2%. (Google)

OpenAI launched Businesses in ChatGPT in beta on November 10, letting companies claim a profile in a new Business Hub and reach the 1.2 billion people who use ChatGPT each week. (OpenAI)

1. Anthropic Launches Claude Fable 5.2 and Mythos 5.2

Anthropic on November 9 launched Claude Fable 5.2 and Claude Mythos 5.2, the next generation of its flagship model line, reports Anthropic. The two names describe the same underlying model with different safeguards. Fable ships with cybersecurity and biology safeguards and is open to all users, while Mythos is restricted to organizations in Anthropic's Cyber and Life Sciences Verification Programs.

The price moved in the direction users like. Fable 5.2 costs $8 per million input tokens and $40 per million output tokens, down from $10 and $50 for Fable 5.1, a 20% cut. An optional fast mode at $16 and $80 per million tokens speeds generation up to 2.5x, and output is up to 30% faster than Fable 5.1. The model keeps a 1M-token context window with up to 128K tokens of maximum output.

The benchmarks, as Anthropic reports them, show a large jump. On FrontierCode 2.0, its agentic coding benchmark, Fable 5.2 scored 49.1%, up from 32.4% for Fable 5.1 and ahead of Opus 5.5 at 45.5%. On Humanity's Last Exam with tools it reached 74.2%, up from 65.6%. Mythos 5.2 scored 89.6% on CVE Rediscovery, against 78.3% for Mythos 5.1. These are vendor-reported figures, so treat them as Anthropic's own claims rather than independent measurements.

The most striking demonstration was not a benchmark. A team of about 1,000 Fable 5.2 agents fully formalized Fermat's Last Theorem in Lean in 30 hours with minimal human intervention. The model also solved two open problems in black hole physics and designed three working electronic devices. Whether or not those feats translate directly into everyday coding work, they show what a large team of coordinated agents can do when the model is good enough to check its own work.

2. Anthropic Kicks Off a Trillion-Dollar IPO Roadshow

The same week, Anthropic opened the roadshow for an initial public offering that would dwarf every listing before it, reports Reuters. The company and some existing shareholders seek to raise up to $64.4 billion at a valuation of about $1 trillion, offering 825 million shares priced between $70 and $78 apiece, with a listing planned on Nasdaq under the ticker ANT.

Demand so far has been intense. The IPO drew about twice the orders for the shares available by the end of Tuesday, the first roadshow day, and by the end of Thursday the book was 6x oversubscribed, reports Reuters. The roadshow was scheduled to run through the following Wednesday, when the company is expected to price the $64 billion deal. Alongside the IPO, $26 billion in investments are lined up from Google, Nvidia, AMD, Broadcom, Micron, Microsoft, and TSMC through a private placement at the IPO price.

The numbers behind the valuation are just as large. Sell-side projections put Anthropic revenue at around $52 to $53 billion in 2026, rising to $135 to $137 billion in 2027, with banks estimating 2028 revenue around $270 to $280 billion. Those are projections, not results, and the same Reuters report notes investors are asking whether Anthropic can hold its AI lead. Still, a trillion-dollar valuation would make Anthropic the most valuable AI company ever to go public, and the biggest test yet of whether the market believes AI revenue can keep compounding at this pace.

Read the earlier chapter of this story in last week's AI roundup, which covered the METR assessment of AI speeding up Anthropic's own research.

3. Google Puts Gemini 4 Hydrogen Into Search

Google on November 12 launched Gemini 4 Hydrogen, the third model in the Gemini 4 family after Argon in September and Carbon in October, reports Google. Hydrogen is built for rapid execution in everyday coding and agentic workflows, and it is rolling out globally across Google AI Studio, the Gemini API, Gemini for business, the Gemini app, Google Search, and Google Antigravity for Pro and Ultra users.

The pricing is deliberately simple: a flat $1.50 per million input tokens and $7.50 per million output tokens across the entire 1 million-token context window. On the Artificial Analysis Intelligence Index, Hydrogen scores 51 against 50 for Argon at the default high thinking level, reaching 54 at the highest xhigh setting.

The comparison that matters most for everyday users is against Gemini 3.8 Flash, the fast model Hydrogen effectively replaces. Hydrogen finishes tasks 2.5x faster at its high default setting and more than triples Terminal-Bench 4.1 scores, 57% versus 18.2%, while cutting fabricated claims by more than 5x. Because Hydrogen now powers Google Search globally, including AI Mode and AI Overviews, those improvements reach every search user, not just developers.

Google also added a new Orchestrator feature in Antigravity that dynamically switches between Carbon and Hydrogen in one session. On Terminal-Bench 4.0 it reached near-Carbon task success while consuming 24% less quota and completing 19% faster. The message is clear: the race is no longer only about the smartest model, but about the smartest model at the right price and speed for each task.

4. OpenAI Turns ChatGPT Into a Business Directory

OpenAI on November 10 launched Businesses in ChatGPT in beta, reports OpenAI. Companies can claim their business in a new Business Hub, manage their information, and see aggregated insights on how people engage with them in ChatGPT, including mentions, website clicks, and discovery topics.

OpenAI frames the move around reach: 1.2 billion people use ChatGPT each week. Business Agents, branded chat assistants inside ChatGPT, are available to eligible ecommerce and online service businesses in the United States, and businesses can also sponsor a Business Agent as a ChatGPT Ads destination. Any eligible business located where ChatGPT is available can claim a profile, and Shopify merchants can connect their store from Business Hub.

There is a careful caveat. Claiming a profile does not guarantee higher placement or sales, and ChatGPT keeps making independent recommendations. No price has been published for claiming a profile or for Business Agents. The direction is unmistakable, though: ChatGPT is becoming a place where people discover businesses, not just ask questions, and OpenAI is building the commercial layer on top of that discovery.

How the Week's Stories Stack Up

Aspect

Story

Key Number

Source

Flagship models

Anthropic launches Claude Fable 5.2

49.1% on FrontierCode 2.0, 20% price cut, 1,000 agents formalized Fermat's Last Theorem in 30 hours

Anthropic

Markets

Anthropic IPO roadshow opens

Up to $64.4 billion at about a $1 trillion valuation, 6x oversubscribed

Reuters

Search and speed

Google launches Gemini 4 Hydrogen

$1.50/$7.50 per million tokens, 57% on Terminal-Bench 4.1, now in Search globally

Google

Commerce

OpenAI launches Businesses in ChatGPT

1.2 billion weekly users, Business Hub in beta

OpenAI

The thread across all four stories: the AI business is growing up fast. A model launch now comes with a trillion-dollar valuation attached, a search upgrade reaches a billion people overnight, and the chatbot itself is becoming a storefront. Last week was about what AI can do and how fast it improves. This week is about what AI is worth, and who gets to sell through it.

Read this article in Indonesian: Berita AI: IPO Triliun Dolar Anthropic, Claude Fable 5.2, dan Gemini 4 Hydrogen di Search

Tags: ai news · artificial intelligence · anthropic · claude fable · gemini 4 hydrogen · openai · chatgpt

All posts