Smartest Mode: Full Thinking Power for Your Agents

From now on you can switch your SMarTrAgents to the most capable AI model available, for the moments a task really needs full thinking power. Your platform data stays on our servers in the EU, while the model itself runs with Anthropic in the US. We say this clearly so you can decide for yourself what to send.

Not every task needs the same model. A quick summary, a fast reply, a standard text, a nimble, affordable model handles that. But a tricky analysis, a well-thought-out plan, a clean piece of code? That’s when you want the very best. That’s exactly what the new Smartest Mode is for.

Three tiers, one decision: you choose how much thinking power

At SMarTrAgents you pick, per chat, how much power your agent gets:

• Normal Mode, fast and efficient for everyday work: summaries, drafts, quick research.
• SMarTr Mode: the stronger tier for more demanding text, code and analysis.
• Smartest Mode, the top class: here your agents work with Claude Opus 5, one of the most intelligent models available. For the moments when every detail counts.

You don’t need to be an AI expert to pick the right tier. One click on the mode, done. And which model is the best choice under the hood is something we re-check every week and set automatically for you. You always get the best available performance, without ever having to worry about model names.

Why Opus 5 makes the difference

Smartest Mode is not a marketing label. There’s real compute behind it: Opus 5 reasons in multiple steps, keeps long tasks coherent, finds bugs other models miss, and plans complex work from start to finish. For your agents that means: hard jobs aren’t just “somehow” done, they’re thought through, thorough and self-checked.

When Smartest Mode pays off:

• a real strategy or financial analysis instead of a keyword answer
• code that should work on the first try
• a long, single-thread task where the agent has to keep the overview
• anything where a mistake would be costly

Being honest: where your data lives, where the model runs

For us, data protection isn’t a footnote, so let’s be precise: your accounts, chats and files live on our servers in the European Union. Model execution in Smartest Mode is currently handled by Anthropic in the US, which is where the contents of your request go for the duration of processing. Details and your rights are in our privacy policy. Especially if you work with sensitive content, you should know this and be able to decide which task goes into which mode.

And we’re going further: our goal is to bring European processing even more into our own hands for our customers. “Made in Germany” isn’t a sticker for us, it’s the direction.

New: “Nobody”, the rabbit that thinks along

Fittingly, our platform has gained a small new helper: Nobody, a white rabbit that’s always there for you in the bottom right. Ask it anything about the cloud, it looks things up in the documentation and on the web, gives you short, clear answers and links you to the matching wiki section. It winks at you, visibly thinks, and pulls out its smartphone when it’s researching for you. Nobody doesn’t execute anything itself, it explains and shows you the way; the rest is done by you or your agents.

Human with machine

Smartest Mode is our answer to a simple belief: human with machine beats both human and machine. You bring the idea, the judgement, the responsibility, the AI brings the thinking power. And with Smartest Mode you can, when it matters, switch on the very best.

Try it: open a chat, switch to Smartest Mode and give your agent your hardest task. Getting started is available from €25.99/month, the full feature set is in the Professional plan at €36.99/month.

Get started at smartragents.ai

Human with machine beats both human and machine.

Imprint · Privacy · Terms · Right of withdrawal