Anthropic dropped a new AI model on July 24, 2026, and it is already doing something strange. Opus 5 costs half as much as the model the internet could not stop talking about in June, yet it matches that older model on the work most people actually do. The result is a quiet changing of the guard. Claude Fable 5, the frontier model everyone called a breakthrough a few weeks ago, is suddenly the one people forgot.
What Opus 5 actually is
Opus 5 is Anthropic’s newest general-purpose model. It arrived on July 24, 2026, priced at $5 per million input tokens and $25 per million output tokens. Compare that with Fable 5, which launched June 9, 2026 at $10 per million input tokens and $50 per million output. That is exactly half the price for the everyday tier.
Opus 5 ships with a 1 million token context window, so it can hold an entire large codebase or a long research document in memory at once. Anthropic positions it as the model you reach for daily, not just for special projects.
Why Fable 5 is being left behind
Here is the part that surprised the industry. On coding and knowledge-work benchmarks, Opus 5 actually outperforms Fable 5, according to both Anthropic and CNBC. So the cheaper model is also the better one on the tests people use to pick a daily driver. For most users, that math is hard to argue with.
Fable 5 still leads where tasks get long and hard. Anthropic’s own product lead, Dianne Penn, told Reuters to pick Opus 5 for value and Fable 5 for “days-long, very autonomous projects.” Fable 5 also stays ahead on risky dual-use work like offensive cybersecurity and biology research, areas where Opus 5 is deliberately held back by tighter safeguards.
The model nobody mentions: Mythos 5
Lost in the Opus versus Fable story is the model that sits above both. Claude Mythos 5 is the same underlying system as Fable 5 but with fewer safety restrictions, and it has the strongest cybersecurity capability of any model Anthropic has shipped. It is reserved for a small group of cyberdefenders and infrastructure partners through a US government program. For the average user, it is out of reach, which is exactly why the Opus 5 versus Fable 5 comparison dominates the conversation.
What this means for you
If you use AI for writing, coding, research, or daily office work, Opus 5 is now the sweet spot. It is cheaper, it holds more context, and it beats the model that was the headline act last month on the work you do most. Fable 5 is not gone, but its job has narrowed to the longest, riskiest, most autonomous tasks. The frontier did not get cheaper by accident. Anthropic is racing a cost-conscious market where enterprises want clear returns before they spend, and Opus 5 is the answer.
The takeaway is simple. The model everyone forgot is still the most capable general model for the hardest jobs. But the model everyone will actually use from here is the one that costs half as much and does almost everything just as well.
