Anthropic made Claude Opus 5 available today, giving its flagship line a new default model that it says is stronger than Opus 4.8 while costing the same. The model now sits at the center of Claude Max and is the strongest option on Claude Pro, putting a fresher system in front of users the moment they sign in.
The pitch is simple and direct: more capability for the same cost. Anthropic says Claude Opus 5 comes close to the frontier intelligence of Claude Fable 5 at half the price, and that advantage shows up where many users will notice it first. On Frontier-Bench v0.1, it surpasses all other models, while on CursorBench 3.2 it reaches within 0.5% of Fable 5’s peak score at max effort, and does so at half the cost per task.
That is why the launch matters now. Claude Opus 5 is the latest test of whether a model can move closer to the top of the field without pushing users into a higher price tier. Anthropic says the model is also the new state-of-the-art on coding and knowledge work evaluations such as Frontier-Bench and GDPval-AA, and that customers can use its effort setting to tilt toward intelligence or save tokens for faster, cheaper results.
There is still one clear line where the new model does not lead. In rigorous evaluations with private-sector and government partners, Claude Opus 5 stays behind Mythos 5 on cybersecurity tasks, even as Anthropic says it did not train the model on cyber work in the way it did with Opus 4.8. The company says the model improved anyway because it became more generally capable, and that puts the launch in a familiar place for frontier AI: ahead in broad use, but not across every hard test.
Anthropic is also leaning on safety as part of the case for the release. During pre-deployment testing, its automated behavioral audit found Claude Opus 5 to be its most aligned model to date, with better adherence to the Constitution, lower rates of deceptive behavior and less susceptibility to misuse than Opus 4.8, Sonnet 5 or Fable 5. The result is a model the company is positioning not just as cheaper and stronger, but as the one it wants people to use every day.
What happens next is less about a future launch than about adoption. Claude Max and Claude Pro users now have a new baseline, and the question is whether the mix of higher performance, lower cost and broader general capability is enough to pull work away from older models fast enough to make Claude Opus 5 the practical default outside Anthropic’s own packaging.
