Anthropic leapfrogs OpenAI as the most valuable AI startup — and drops a new model (www.businessinsider.com)

Anthropic has secured a monumental $65 billion fundraising round, elevating its valuation to $965 billion and positioning it as the most valuable AI startup, overtaking OpenAI. On the same day, the company unveiled Claude Opus 4.8, an iteration of its general-purpose large language model, boasting enhanced capabilities in coding, reasoning, and general knowledge tasks. This development is particularly significant as it underlines the intensifying competition between AI labs, both of which are racing towards IPOs and striving to capture a larger share of the AI market by selling advanced coding tools to developers. The new Opus 4.8 model introduces an "effort control" feature, allowing users to manage the computational resources allocated to responses, thereby facilitating more efficient budgeting for AI usage. Notably, the fast mode of Opus 4.8 is designed to be three times more cost-effective than previous versions, appealing to companies reassessing their AI expenditures. While this release provides incremental improvements, it is positioned as "weaker" compared to the anticipated Mythos model, which is currently on hold due to cybersecurity concerns. Anthropic is actively developing safeguards for Mythos and aims to deploy it soon, signaling a strategic push into the critical cybersecurity domain.