Anthropic launched Claude Opus 5 on July 24, positioning it as a thoughtful, proactive AI that comes close to the frontier intelligence of Claude Fable 5 at half the price. The new model is priced at $5 per million input tokens and $25 per million output tokens—the same cost as Opus 4.8, but with notably stronger performance.

The release marks Anthropic’s fourth Claude 5 model in less than two months. The company has shifted strategy from blockbuster launches to rapid incremental improvements on capability, cost, and speed. Each new version narrows the gap between frontier and mainstream models.

Where Opus 5 Excels

On coding and knowledge work, Opus 5 is now the state-of-the-art. Benchmarks like Frontier-Bench and GDPval-AA show Opus 5 leading against previous generations. It remains behind Mythos 5 on cybersecurity tasks, but for most professionals, the gap is negligible.

Opus 5 is now the default model on Claude Max and the strongest option on Claude Pro. This signals that Anthropic believes the model is ready for everyday use by paying customers, not just for specialized workloads.

The Price War Intensifies

The timing matters. OpenAI just cleared GPT-5.6 for release, and Chinese models like Moonshot’s Kimi are gaining traction for cost-sensitive workloads. By holding Opus 5 at the same price as a weaker predecessor, Anthropic is betting that performance gains will drive adoption without needing to undercut competitors on price.

For builders, the math is simple. If Opus 5 solves your problem and costs half what Fable does, you switch. The burden is now on OpenAI and others to show why their models deserve premium pricing.

The real story: AI pricing is collapsing as models improve. In six months, “frontier” capability will be a commodity, and builders will shop on speed and reliability instead.