Anthropic Doubles Claude Code Limits After SpaceX Compute Deal
Anthropic doubled Claude Code limits for paid plans, removed Pro/Max peak-hour reductions, and raised Opus API rate limits without raising subscriptions.
By AI Pricing Guru Editorial Team
AI Pricing Guru articles are maintained by the editorial workflow behind the site: daily pricing snapshots, provider source checks, and review passes for model launches, subscription limits, and billing changes.
Subscription pricing is where teams often fool themselves. A $20 plan can be perfect for one person and still be the wrong benchmark for a product feature serving thousands of users.
Anthropic just made Claude subscriptions materially more valuable for developers, without announcing a subscription price increase.
The short version: Claude Code’s five-hour rate limits are now doubled for Pro, Max, Team, and seat-based Enterprise plans. Anthropic also removed the peak-hours limit reduction for Claude Code on Pro and Max accounts, and raised Claude Opus API rate limits across usage tiers.
This is worth paying attention to because it isn’t a normal model launch. It’s a capacity-driven pricing/value change: same headline subscription prices, more usable Claude Code throughput.
What changed
Anthropic says three changes are effective immediately:
| Area | Before | Now | Pricing impact |
|---|---|---|---|
| Claude Code on paid subscriptions | Lower five-hour rate limits | Five-hour limits doubled for Pro, Max, Team, and seat-based Enterprise | More developer usage at the same plan price |
| Peak-hour behavior | Pro and Max could see reduced Claude Code limits during peak hours | Peak-hour limit reduction removed for Pro and Max | More predictable workday usage |
| Claude API | Lower Opus rate limits by tier | Claude Opus 4.x rate limits raised considerably | Better throughput, same token prices |
| Subscription prices | Claude Pro $20/mo; Max tiers at $100/mo and $200/mo in our tracked data | No announced price increase | Effective value-per-dollar improves |
The key nuance: Anthropic’s announcement is specifically about Claude Code limits, not a blanket promise that every Claude chat limit doubled across every product surface.
For normal Claude app chat, keep treating the published subscription limits as dynamic and account-dependent. For developers using Claude Code heavily, this is the meaningful change.
Why this happened
The change is tied to a new compute partnership with SpaceX. Anthropic says it will use all compute capacity at SpaceX’s Colossus 1 data center, adding more than 300 megawatts of capacity and over 220,000 NVIDIA GPUs within the month.
Anthropic frames the deal as a near-term capacity unlock for paid Claude users, especially Claude Pro and Claude Max subscribers. It also sits on top of the company’s larger compute commitments with Amazon, Google/Broadcom, Microsoft/NVIDIA, and Fluidstack.
In plain English: Claude has been capacity-constrained, especially for heavy coding and agent workflows. Anthropic bought more capacity, and the first visible customer-facing result is higher limits.
Why this matters for Claude Code users
Claude Code is exactly the kind of workload that exposes subscription limits quickly. A coding session can include large repositories, repeated tool calls, long diff reviews, tests, retries, and follow-up prompts. That burns through quota much faster than casual chat.
Doubling the five-hour window changes the practical math:
- Claude Pro becomes more viable for real coding sessions instead of occasional help.
- Claude Max 5x is easier to justify for developers who previously hit limits before the end of a work block.
- Claude Max 20x becomes closer to an all-day coding-agent plan for many individual power users.
- Team and Enterprise seats get a productivity bump without needing an immediate plan migration.
The peak-hours change may matter even more than the doubling. If you use Claude Code during a normal workday, removing peak-hour reductions should make limits feel less random.
API buyers also get a throughput upgrade
Anthropic also updated its API rate-limit table for Claude Opus 4.x. Current published standard-tier Opus 4.x limits are:
| API tier | RPM | Input tokens/min | Output tokens/min |
|---|---|---|---|
| Tier 1 | 50 | 30,000 | 8,000 |
| Tier 2 | 1,000 | 450,000 | 90,000 |
| Tier 3 | 2,000 | 800,000 | 160,000 |
| Tier 4 | 4,000 | 2,000,000 | 400,000 |
This doesn’t change Anthropic’s per-token API prices by itself. Opus is still a premium model, so higher throughput can mean faster spend if you don’t have routing and budgets in place.
But for teams that were rate-limit blocked rather than budget blocked, the practical effect is positive: more room for batch analysis, agentic coding, evaluation runs, and high-value reasoning workloads.
Is Claude Pro now a better deal?
Yes, for developers who use Claude Code directly.
At the same $20/month headline price, Claude Pro now includes more Claude Code capacity and less peak-hour friction. That makes it more competitive against ChatGPT Plus, Cursor, GitHub Copilot, and other developer subscriptions.
The caveat is billing surface. This change helps first-party Claude Code subscription usage. It doesn’t undo Anthropic’s earlier distinction between first-party Claude usage and some third-party harness usage, which may still require separate Extra Usage or API billing depending on the setup.
So the practical recommendation is:
- Use Claude Pro if you want affordable individual Claude Code access and only sometimes hit limits.
- Use Claude Max 5x if Claude Code is part of your daily development workflow.
- Use Claude Max 20x if Claude is your primary coding agent and interruptions are costly.
- Use the Anthropic API for production software, automation, customer-facing workflows, or anything that needs token-level cost attribution.
What teams should do now
- Re-test Claude Code limits this week. If you previously downgraded or avoided Claude Code because of five-hour caps, the product may now feel meaningfully different.
- Don’t assume chat limits doubled. The official change is about Claude Code’s five-hour limits and API Opus rate limits.
- Keep budget alerts on API workloads. Higher Opus throughput is useful, but it can also let expensive jobs run faster.
- Compare subscription vs API usage honestly. Human developer workflows belong on subscriptions; production workflows belong on metered API billing.
- Watch for follow-on plan changes. More capacity gives Anthropic room to make Max, Team, and Enterprise more aggressive without cutting token prices.
My read
This is a real article-worthy pricing story because it changes the effective value of Claude subscriptions for developers.
Anthropic did not announce cheaper API tokens or lower plan prices. Instead, it improved the denominator: more Claude Code usage for the same subscription price and higher Opus API throughput for teams that were hitting rate limits.
For Claude Code users, that’s a direct upgrade. For API buyers, it’s a capacity win, not a cost discount.
For broader context, see our Anthropic API pricing guide, Claude subscription comparison, and the earlier billing update on OpenClaw usage under Claude plans.
Sources: Anthropic: Higher usage limits for Claude and a compute deal with SpaceX, Anthropic API rate limits, and AI Pricing Guru’s tracked Claude subscription pricing data.