Devin Rate Limit Exceeded: What It Means and How to Fix It

When Devin hits a rate limit, it stops mid-task and you get a "rate limit exceeded" error. This happens when you've exhausted your monthly ACU (Agent Compute Unit) allocation or hit the concurrent-task cap for your plan. The fix depends on which limit you've hit: daily task throttling, monthly ACU cap, or team seat concurrency.

What are ACUs and why do they cause rate limits?

Devin bills by ACU (Agent Compute Unit), not by message or hour. Each task Devin runs consumes ACUs based on the complexity and duration of the work. According to Devin's pricing page, plans come with a fixed ACU budget per billing cycle. When that budget runs out, new tasks are blocked until it resets or you top up.

• Starter plan: limited ACU allocation, single concurrent session
• Teams plan: higher ACU pool shared across seats, multiple concurrent sessions
• Enterprise: custom ACU limits negotiated per contract

Rate limits on Devin are almost always about ACU exhaustion or concurrent session caps, not per-minute request throttling like a traditional API.

What triggers a Devin rate limit exceeded error?

There are three distinct scenarios that produce a rate limit error in Devin:

1. Monthly ACU cap hit: You've used all allocated compute for the billing period. Devin won't start new sessions until the cycle resets or you purchase more ACUs.
2. Concurrent session limit: Your plan allows only a fixed number of parallel Devin sessions. Launching more than the plan allows triggers an immediate rate limit response.
3. Underlying model limits: Devin orchestrates sub-agents and calls models like Claude or GPT-4. If those upstream providers are throttling, Devin inherits the error and surfaces it as a rate limit to you.

How do you check your remaining Devin ACUs?

Log into the Devin dashboard and navigate to Settings > Usage. The usage panel shows your current ACU consumption, remaining balance, and billing cycle reset date. There is no in-IDE indicator; you have to check the web dashboard manually. Unlike tools like Claude Code, Devin doesn't have a built-in slash command to query limits mid-session.

How to fix Devin rate limit exceeded

Depending on which limit you've hit, the options differ:

If you hit the monthly ACU cap

• Wait for your billing cycle to reset (check the reset date in Settings > Usage)
• Purchase additional ACUs as an add-on top-up through the Devin dashboard
• Upgrade to a higher plan tier with a larger ACU allocation

If you hit the concurrent session limit

• Wait for an active session to complete before launching a new one
• Upgrade to Teams or Enterprise for higher concurrency
• Queue tasks sequentially rather than kicking off multiple sessions at once

If the error is upstream (model throttling)

• Retry after a few minutes. Upstream provider limits are usually short-lived
• Check Devin's status page for ongoing incidents

Devin plan limits compared

Plan	ACU Allocation	Concurrent Sessions	Overage Option
Starter	~250 ACUs/month	1	Top-up available
Teams	Shared pool (varies by seats)	Multiple	Top-up available
Enterprise	Custom	Custom	Negotiated

As of July 2026, Devin's exact ACU-to-task conversion varies by task complexity. Cognition's documentation notes that a typical agentic task (write, test, and PR a feature) consumes roughly 10-40 ACUs depending on scope.

How do you reduce ACU consumption on Devin?

You can slow ACU burn by scoping tasks more narrowly. Instead of asking Devin to "build a user authentication system," break it into smaller, bounded tasks. Each session startup has overhead cost, but sprawling open-ended sessions burn far more ACUs. Give Devin clear acceptance criteria so it stops when done rather than continuing to explore.

• Use specific, bounded task descriptions with clear done conditions
• Review and close idle sessions rather than leaving them open
• Monitor the usage dashboard mid-cycle, not just at the end
• Avoid re-running identical tasks: Devin doesn't cache, so duplicate runs consume fresh ACUs

How does Devin's rate limiting compare to other AI coding tools?

Devin's ACU model is different from message-based tools. Cursor, Copilot, and Claude Code all gate on request counts or token budgets per period. Devin gates on compute output, meaning a single complex task can exhaust a significant chunk of your monthly budget in one go. If you're used to Claude Code's 5-hour rolling window (see Claude Code's 5-hour vs weekly limit explained), Devin's monthly lump allocation feels less predictable.

For comparison, OpenAI Codex usage limits reset on a different cadence and don't use the ACU model at all. And unlike Claude Code rate limits, Devin doesn't offer a built-in way to monitor your remaining budget from within the editor itself.

If you work across multiple AI coding tools and want live visibility into Claude Code's 5-hour window and weekly cap alongside your other usage, Usagebar tracks Claude Code limits in your macOS menu bar with notifications at 50, 75, and 90% usage, one-time $9 purchase.

When does Devin's rate limit reset?

Devin resets ACU allocations on your monthly billing anniversary, not on a calendar month basis. If you subscribed on the 15th, your ACUs refresh on the 15th of each month. The exact reset date is visible in the dashboard under Settings > Usage. There's no weekly rolling window; it's a hard monthly reset.

Key takeaways

• Devin rate limits come from ACU exhaustion, concurrency caps, or upstream model throttling
• Check your remaining ACUs at app.devin.ai under Settings > Usage
• Purchase ACU top-ups or upgrade plans to unblock immediately
• Scope tasks narrowly to reduce ACU burn per session
• ACUs reset on your billing anniversary, not on the first of the month
• For other tools like Claude Code, see how Claude Code rate limits work and Devin vs Copilot for a broader comparison

Sources

• Devin pricing and ACU plans – devin.ai
• Devin dashboard usage panel – app.devin.ai
• Devin status page – status.devin.ai