Anthropic Surpasses OpenAI with Claude Opus 4.8

Anthropic Surpasses OpenAI with Record Valuation

Anthropic recently announced an impressive fundraising round of $65 billion, bringing its total valuation to $965 billion. This new position places Anthropic ahead of OpenAI as the most valuable AI startup, as both companies prepare to go public. Anthropic's latest valuation represents a massive leap from its $380 billion valuation in February.

The newly launched Claude Opus 4.8 is touted as a significant improvement over its predecessors. It particularly excels in coding, reasoning, and general knowledge. This advancement is part of a broader strategy where Anthropic and OpenAI compete to dominate the tech market and attract developers. OpenAI's latest fundraising round in March valued the company at $852 billion, and its latest model, GPT-5.5, was launched in April.

Claude Opus 4.8: Innovations and Savings

The Opus 4.8 model introduces a "effort control" system, allowing users to manage the computing power allocated to each task. This feature is crucial for companies looking to optimize their AI spending. Additionally, Opus 4.8's fast mode is offered at a cost three times lower than previous versions, providing an economical solution for clients.

However, Opus 4.8 is not the much-anticipated Mythos model, which remains on hold due to cybersecurity concerns. Mythos had generated interest for its ability to detect vulnerabilities, but Anthropic is currently working with trusted partners to develop adequate security measures.

Cybersecurity Challenges and Anthropic's Future

Although Opus 4.8 is less advanced than Mythos in terms of cybersecurity capabilities, it still presents intriguing features. For instance, in tests, Opus 4.8 could be deceived into presenting itself as human in about 3% of cases. This limitation highlights the ongoing challenges in developing robust AI models.

Anthropic continues to collaborate with cybersecurity experts to ensure that its future models, like Mythos, can be deployed safely. This cautious approach may well influence the future of AI, balancing innovation and security.