Sliq Logo Sliq

Perplexity Computer Pricing & Credits Explained (2026)

Perplexity Computer launched on February 25, 2026, and the headline number is simple: $200 per month. That's the price of Perplexity Max, the subscription tier required to access Computer.

But $200/month is the starting point, not the whole picture. Computer runs on a credit system, and the mechanics of that system - how fast credits burn, what happens when they run out, and how high your bill can actually climb - are worth understanding before you commit. This isn't a warning. Perplexity has built in real spending controls. But the difference between using Computer informed and using it blind is potentially hundreds of dollars a month.

The subscription: $200/month for Max

Perplexity Computer is not a standalone product. It's a feature of Perplexity Max, the $200/month tier that also includes unlimited Pro searches, access to advanced models (GPT-5.2, Claude Opus 4.6, Gemini), Sora 2 Pro video generation, the Comet AI browser, and Labs for building dashboards and apps.

If you're already on Max for other reasons, Computer is effectively a free addition. If you're signing up specifically for Computer, you're paying $200/month for the whole suite whether you use the rest of it or not.

There's no way to access Computer on the Pro plan ($20/month). Perplexity has said Pro and Enterprise access is "on the roadmap," but no timeline has been announced. If you're a Pro subscriber and interested in Computer, upgrading to Max is currently the only path.

The credit system: 10,000 per month, then pay as you go

This is where the pricing gets interesting.

Max subscribers receive 10,000 credits per month included with their subscription. Every Computer task consumes credits based on the complexity of the work - how many sub-agents are spawned, which models handle each step, and how many iterations the orchestrator runs before delivering a result.

Here's the catch: Perplexity has not published a per-task credit conversion table. There's no page that says "a research task costs X credits" or "building an app costs Y credits." Credit consumption varies, and the only way to see what a task cost is to check after it runs - either through the overflow menu on any thread or at perplexity.ai/account/usage.

This makes budgeting genuinely difficult. One real estate industry blogger put it well: if a complex due diligence workflow burns 500 credits, you're running 20 of those per month on your base plan. If it burns 2,000, you're running 5. Without published rates, you're estimating by feel until you build up enough usage history to see your own patterns.

For context, Perplexity's help center describes credit consumption as depending on "the complexity and resources required for each task." That's accurate but not especially helpful when you're trying to figure out if 10,000 credits will last the month.

The launch bonus: generous, but temporary

New Max subscribers currently receive a one-time bonus of 35,000 credits on top of their 10,000 monthly allocation. That's 45,000 credits in your first month - four and a half times the ongoing monthly amount. For existing Max users, the bonus was granted at launch.

This is genuinely generous, and it gives you a real runway to test workflows and understand your usage patterns. The important detail: the bonus credits expire 30 days after they're granted. They're consumed first (before your monthly credits), so you'll naturally burn through them early.

The practical implication is that your first month with Computer will feel more abundant than your second month. That's fine as long as you're aware of it. Use month one to experiment, track how many credits different types of tasks consume, and calibrate expectations for what 10,000 credits per month actually looks like for your specific usage.

Auto-refill and spending caps: the guardrails

When your monthly credits run out, you have two options: stop using Computer until next month, or buy more credits.

Perplexity offers an auto-refill feature that automatically purchases credits when your balance drops below 500 (roughly $5 worth). Auto-refill is off by default - you will never be charged for additional credits without explicitly opting in. That's an important design choice that protects users from surprise bills.

If you do enable auto-refill, you set two limits: a refill amount (how many credits to purchase each time) and a monthly auto-refill limit (the maximum you'll spend on auto-refill per month). The default monthly limit is $200.

On top of that, every Max account has a monthly spending cap that limits total credit charges per billing period. The default cap is $200. You can lower it or raise it up to $2,000.

So the math on your theoretical maximum monthly spend:

  • Subscription: $200
  • Additional credits (up to spending cap): $200 default, $2,000 maximum
  • Monthly range: $200 to $2,200

Most users won't get anywhere near $2,200. But if you're running complex workflows continuously and have raised your spending cap, it's technically possible. One Reddit user estimated that heavy Computer usage could run to $1,500/month in credits alone - on top of the $200 subscription. That's the high end, not the typical case, but it shows the range.

What happens when credits run out

If your credits hit zero and auto-refill is off (or you've hit your spending cap), two things happen: active tasks pause, and new tasks are blocked.

The good news is that tasks pause rather than cancel. Your progress is saved, and work resumes automatically once credits become available - whether that's through a manual purchase, an auto-refill, or your next monthly credit refresh. Nothing is lost.

The less-good news is that if you've built workflows that depend on Computer running continuously - monitoring tasks, scheduled reports, long-running research projects - a credit pause interrupts all of it. For casual use this doesn't matter. For anyone treating Computer as always-on infrastructure, the credit ceiling creates a reliability question that flat-rate alternatives don't have.

The model selection lever

Computer orchestrates 19 models, routing each sub-task to whichever model handles it best. Claude Opus 4.6 runs the core reasoning engine. Gemini handles deep research. GPT-5.2 manages long-context recall. Grok handles lightweight tasks. Nano Banana does images. Veo 3.1 does video.

By default, the orchestrator picks the best model for each step. But you can also manually assign models to specific sub-agent tasks. This matters for cost because different models consume credits at different rates. Routing a lightweight task to Grok instead of Claude Opus uses fewer resources and fewer credits.

Most users will let the orchestrator decide, and that's usually the right call - it's the whole point of the multi-model architecture. But if you're watching your credit consumption and want to optimize, model selection is the main lever you have. It's worth experimenting with during your bonus credit period to see how much it affects your burn rate.

How it compares to alternatives

Every AI agent has a different cost structure, and each comes with a different type of financial risk.

Claude Cowork: $20-200/month, flat rate. You pay for a Pro ($20/month) or Max ($100-200/month) subscription and that's it. No per-task charges, no credits, no overages. You might hit usage limits on message volume, but you'll never get an unexpected bill. The tradeoff is that you're limited to one model (Claude) and desktop-only workflows. Full comparison here.

OpenClaw: Free software + $30-200+/month in variable costs. The software is free, but you pay for hosting ($5-24/month), API fees to whatever model provider you choose ($15-50/month for moderate use, much higher for heavy use), and your own time for setup and maintenance. The financial risk is API cost spikes - one user burned $3,600 in a single month, and there's no built-in spending cap unless you set one with your API provider. Full cost breakdown here.

Perplexity Computer: $200/month + variable credit costs. You get a capable multi-model system with built-in spending controls. The 10,000 included credits are enough for moderate usage. Heavy usage pushes costs higher, but the spending cap means you always know your maximum. The main risk isn't a surprise bill - it's running out of credits mid-workflow and having everything pause.

The predictability ranking, from most to least predictable: Claude Cowork (flat rate) -> Perplexity Computer (base + capped variable) -> OpenClaw (fully variable, no built-in cap).

Who the pricing actually works for

Perplexity's own executives have described Computer as targeting people making "GDP-moving decisions." That's marketing language, but the pricing structure backs it up. At $200/month minimum, Computer is priced for professionals whose time savings justify the cost - consultants, analysts, founders, anyone regularly doing multi-step research and execution work that would otherwise take hours.

If you're doing a few research tasks a week, 10,000 credits will likely last the month and your total cost stays at $200. That's competitive with other professional AI tools, especially given the multi-model orchestration and 400+ integrations.

If you're running Computer as daily infrastructure - continuous monitoring, frequent complex projects, long-running autonomous workflows - you should expect to go beyond the included credits. Budget $300-500/month as a realistic range for heavy professional use, and set your spending cap accordingly.

If you're curious but not sure whether you'd use it enough to justify $200/month, the 35,000 bonus credits give you a meaningful trial period. Sign up, use it hard for a month, track your credit consumption, and make the decision with real data instead of estimates.

The bottom line

Perplexity Computer's pricing is more thoughtful than it first appears. The $200/month subscription gives you a capable multi-model agent with real spending controls - auto-refill off by default, configurable caps, and tasks that pause rather than rack up charges. It's a system designed to let you scale up intentionally rather than get caught off guard.

The gap is transparency on credit consumption. Until Perplexity publishes clearer guidance on how many credits different types of tasks consume, every new user is essentially running their own pricing experiment during their first month. That's manageable - especially with the launch bonus - but it means your month-one job is to understand your own usage before committing to month two.

For most professional users who stay within the included credits, the total cost is $200/month and nothing more. For power users who push beyond that, the spending cap ensures you're always in control of the ceiling. Both of those are good outcomes. The pricing works - you just need to understand the mechanics before you start.


This is part of a series on AI agents in 2026. See also: Perplexity Computer vs Claude Cowork, Perplexity Computer vs OpenClaw, How Much Does OpenClaw Actually Cost?, and Best OpenClaw Alternatives That Don't Require Coding.

Last updated: March 2026

Multiply yourself with Sliq

Sliq connects to all your tools and can do anything you can - just ask it in Slack.

Try Sliq Free