Anthropic frontier models, cyber capability, funding and safety scrutiny
Key Questions
What funding did Anthropic recently raise?
Anthropic raised $65B at a $965B valuation and filed an S-1 for IPO. Google contributed a $40B investment as part of the round.
What happened to Fable 5 after its release?
Fable 5 was released publicly with safeguards but then disabled following a White House security directive. It shifted to API-only pricing before the shutdown.
How does Opus 4.8 compare to other models?
Opus 4.8 includes effort controls but trails GPT-5.5 in benchmarks. It is now available via AWS Bedrock.
What vulnerabilities did Mythos discover?
Mythos identified over 10,000 vulnerabilities in various systems. A leaked version called Mythos 5 has also surfaced.
How much code does Claude write at Anthropic?
Claude writes over 80% of Anthropic's internal code. Claude Code also rewrote the Bun runtime in just 11 days.
What partnerships or expansions is Anthropic pursuing?
Anthropic opened a Seoul office and saw Microsoft drop Claude Code usage. Salesforce spent $300M on tokens while IBM launched a $5B cybersecurity initiative.
Why did Anthropic apologize regarding guardrails?
The company apologized for invisible guardrails in Claude Fable that affected user experience. This followed reports of overly restrictive or hidden limitations.
What is the current profitability status of Anthropic?
Anthropic reported its first profitable quarter with $10.9B in revenue. An ongoing AI price war is adding pressure on margins.
Anthropic’s reported $65B financing, $965B valuation, IPO filing, Opus 4.8 improvements, and Claude Code adoption reinforce commercial momentum. Mythos/Fable cyber capabilities, Fable 5’s reported disablement after a White House directive, opaque-guardrail criticism, and AI-assisted attacks keep safety and government oversight at peak importance.