Claude Opus 5 is here: near top-tier power at half the price

Anthropic just released Claude Opus 5, and the most interesting news is not a new top score on a benchmark. It is the price. Opus 5 lands close to Anthropic's most capable model, Fable, at roughly half the cost per task. We run Claude in our own marketing engine every single day, so here is the short, practical version: what is new, the difference between Opus and Fable, and which one you should actually use.

In short: what is new in Claude Opus 5?

Claude Opus 5 is Anthropic's new Opus model, launched on 24 July 2026. It comes close to the top model Fable in capability, costs roughly half as much per task, and lets you choose how much effort (and money) it should spend on each task.

The concrete upgrades: a model that thinks by default, that checks its own work and fixes mistakes without you having to push it, and that is now the default model on Claude Max and the strongest option on Claude Pro. The context window is 1 million tokens, the same as the previous Opus generation. It is available in Norway from day one, both in the Claude apps and via the API. So this is not a "coming soon" story. You can start using it today.

The difference between Opus and Fable, explained simply

Fable is Anthropic's most capable model, built for the heaviest and most complex tasks, the kind that can run for hours and days. Opus 5 is nearly as strong at most things, but costs half as much. For most jobs, Opus is more than enough. You bring in Fable when the task truly calls for it.

(Prices are in USD per million tokens, meaning per large volumes of text in and out of the model.)

Fable is the specialist

Fable 5, launched on 9 June 2026, sits at the top of Anthropic's ladder. It is built for autonomous, long-running tasks that would otherwise take a person hours or days: large codebases, heavy research, work that has to hold the thread across a lot of context. It is the most expensive one, and that is the point. You are paying for specialist depth on the genuinely hard lifts.

Opus is the workhorse

Opus 5 closes most of the gap to Fable at half the price. On independent intelligence measures (Artificial Analysis among them) Opus 5 actually comes out above Fable. Even so, Anthropic is clear on one point: Opus 5 is not more capable than Fable overall. Fable keeps the crown on the very heaviest tasks. But for the large majority of jobs you will not notice the difference in practice, and then half the price is an easy call.

The new switch: choose cost or capability per task

Opus 5 has an effort switch with five levels: low, medium, high, xhigh and max. The default is high. You decide how hard the model works, and therefore how much it costs, on each individual task. Simple things run cheaply, heavy things get full power.

For a small team on a budget, this is the most practical piece of the whole launch. A quick rewrite of an email does not need the same engine as a full campaign analysis. Now you can use the same model for both, and only turn up the effort when the task deserves it. There is also a new beta that automatically routes to a lighter model when a request trips the safety filter, so you rarely get stuck.

What does Claude Opus 5 mean for Norwegian businesses?

For Norwegian businesses, Opus 5 means that serious AI use gets cheaper. You get near top-tier power at half the price, and you can dial the cost up and down per task. That makes it easier for a small team to use AI on real work without the bill running away.

Since the model is available in Norway now, this is something you can put into production this week. Our own rule of thumb when we set up AI in the marketing engine: use Opus as the daily default, drop to medium or low on simple tasks, and raise it to xhigh when the task is heavy. You only bring in Fable when you actually hit the ceiling. Cost discipline per task is exactly the kind of move that decides whether AI becomes a cheap workhorse or an expensive habit. It ties closely to the agentic AI we already work with every day, where the model does not just answer, it does the job.

Which Claude model should you choose?

Start with Opus 5 and adjust from there. A simple guide:

• Daily marketing work, writing and analysis, most of what you do: Opus 5 on medium.

• Heavy, long tasks that have to hold the thread: Opus 5 on xhigh or max, or Fable if you hit the ceiling.

• Simple, repetitive tasks: a lighter model such as Sonnet or Haiku does the job fine, and costs less.

This is the same logic as when OpenAI switched to letting you choose level and price instead of a version number. The point is not to always run the most expensive model. It is to match the right model to the task in front of you.

Conclusion

The short verdict: Opus 5 gives you most of the top-tier power at half the price, and lets you control what each task costs. For most Norwegian businesses this is the new default model, with Fable held in reserve for the heaviest lifts.

Want an AI setup that picks the right model for the right task automatically, instead of leaving you to guess? That is exactly what we build in the AI Marketing Engine. Talk to us, and we will show you how it looks in practice.

Frequently asked questions

What is Claude Opus 5?

Claude Opus 5 is Anthropic's new Opus model, launched on 24 July 2026. It comes close to the top model Fable in capability at roughly half the price, has a context window of 1 million tokens, checks and corrects its own work, and lets you choose the effort level across five steps (low, medium, high, xhigh or max) per task.

What is the difference between Claude Opus 5 and Fable?

Fable is Anthropic's most capable model, built for the heaviest and longest tasks, and it is the most expensive. Opus 5 is nearly as strong at most things, but costs half as much per token. Anthropic is clear that Fable is still the most capable on the very hardest tasks, while Opus is the daily workhorse.

What does Claude Opus 5 cost?

Opus 5 costs 5 USD per million input tokens and 25 USD per million output tokens, roughly half of Fable 5, which sits at 10 and 50 USD. You can also lower the cost further by choosing a lower effort level on simple tasks.

Is Claude Opus 5 available in Norway?

Yes. Opus 5 is available in Norway and the EEA from launch, both in the Claude apps and via the API. You can start using it today.

Which Claude model should a small business use?

For most people, Opus 5 on medium covers daily work. Raise it to xhigh on heavy tasks, and use a lighter model such as Sonnet or Haiku for simple, repetitive things. You only bring in Fable when a task genuinely requires the most capable model.