Anthropic has released Claude Opus 5, its newest flagship AI model, and is pairing the launch with an aggressive pitch: better performance than the competition, for less money.
According to a report from finance.biggo.com, Anthropic claims Opus 5 beats OpenAI's GPT-5.6 while costing half as much. A separate write-up from trendingtopics.eu frames the comparison against Anthropic's own lineup, reporting that Opus 5 ships at half the price of Fable 5 — and performs better.
Both of those framings point in the same direction: the headline feature here is price-performance, not just raw capability. That is a notable shift in how frontier AI models get marketed. For the past few years, each new flagship arrived with a premium price tag attached, on the theory that the smartest model was worth paying up for. Anthropic is instead leading with the claim that its most capable model is also the cheaper option.
A caveat worth keeping in mind: the performance claims described in these reports come from Anthropic itself. Vendor-run comparisons against a rival's model are a standard part of AI launches, and independent evaluations typically take days or weeks to catch up. Neither source item includes third-party verification, and no benchmark figures are given beyond the pricing comparison.
Still, the pricing element is concrete in a way benchmark claims usually are not. Cost per token is the single biggest constraint on what companies can afford to build with these models — it determines whether an AI feature can be shipped to millions of users or only to a paying few.
Why it matters: if a leading AI lab can genuinely cut the price of its best model in half, the economics of building AI products get meaningfully easier for everyone downstream — and its competitors face pressure to follow.