Manus rolls out paid subscriptions and an iOS app
Viral AI agent platform Manus just introduced paid subscriptions starting at $39 a month, alongside a brand-new iOS app. Paying users get dedicated task credits and parallel runs, while the service also upgraded its backend model to Anthropic's Claude 3.7 Sonnet.

Manus AI, the viral agent tool that took over timelines just weeks ago, is officially asking users to pull out their wallets.
On Monday morning, the beta platform rolled out two paid tiers starting at $39 a month. Alongside the paywall, Manus launched an iOS mobile app and upgraded the model powering its back end to Anthropic’s Claude 3.7 Sonnet.
Two tiers and an iOS app
The new subscription plans are divided by credits and concurrent task limits.
The cheaper $39-a-month tier comes with 3,900 credits and lets you run two tasks at the same time.
Power users can step up to the $199-a-month plan, which delivers 19,900 credits, five simultaneous runs, and priority access during peak hours. If you run out of credits, top-up packs are available, though pricing has not been disclosed yet.
In a post on X, Manus said it had to temporarily restrict access while working around the clock to scale up its infrastructure. Meanwhile, TechCrunch noted in hands-on tests that while Manus can tackle jobs ranging from wedding invitation sites to baseball score sheets, its actual performance falls short of its loftier marketing promises.
Why it matters
Autonomous agents burn through tokens at an eye-watering pace because they browse, write code, and iterate in loops. In our view, charging real money immediately was Manus’s only way to survive crushing server costs.
At $39 to $199 every month, this is not an impulse purchase. Unless you have specific, multi-step web workflows you urgently want to automate, there is little reason to jump in before the product leaves beta and proves it can consistently deliver on its pitch.
For the AI space, the quick pivot to Claude 3.7 Sonnet highlights just how dependent third-party agents remain on cutting-edge reasoning models. Whether consumers and solo workers will actually stomach enterprise-level subscription rates for agentic tools is a test everyone will be watching closely.
Sources (1)
- techcrunch.com(opens in a new tab)Originaltechcrunch.com