What are Claude Code token limits and how do they work?

Claude Code token limits operate on a 5-hour rolling window that starts with your first message in a session. Pro users receive approximately 44,000 tokens per window, Max5 users get around 88,000 tokens, and Max20 users receive roughly 220,000 tokens. These limits reset every 5 hours. Starting in August 2025, Anthropic introduced weekly limits on top of the 5-hour windows to address unsustainable resource consumption by some users. Model selection impacts usage: Opus 4.5 has higher per-token costs (about 1.7× Sonnet 4.5) and tighter weekly hour caps. Heavy use of Opus will exhaust allocations faster than Sonnet-only usage. Features like 'Explore agents' and 'Plan agents' can burn through tokens rapidly. The Claude Code API provides visibility into estimated cost, tokens used over time, tokens by model, and usage patterns.

How much does Claude Code typically cost per developer?

The average cost for Claude Code is approximately $6 per developer per day, with 90% of users staying below $12 per day. For team deployments using the API with Sonnet 4.5, organizations can expect roughly $100–$200 per developer per month, though actual costs vary based on usage intensity and whether developers run multiple instances. For more cost details, see our blog post on Claude Code token limits.

What cost metrics are important to monitor when using Claude Code?

Key cost metrics to monitor when using Claude Code include: total tokens used by model (e.g., Sonnet vs. Opus) to ensure developers are selecting the most cost-effective model for their tasks; estimated cost over time to identify trends and anomalies; and average estimated cost per commit to assess efficiency and detect potential issues with prompting or workflow configuration. Monitoring these metrics helps organizations spot optimization opportunities and control costs. Faros AI provides visualizations such as average estimated cost per commit.

Where can I find information about Anthropic's new rate limits for Claude Code?

You can read about Anthropic's introduction of new rate limits for Claude Code in this TechCrunch article: Anthropic unveils new rate limits to curb Claude Code power users.

What governance features are available for managing Claude Code usage and costs?

Anthropic's enterprise features for Claude Code include granular spend controls at the organization and individual user level, managed policy settings for tool permissions and file access, and built-in usage analytics. These controls should be used proactively to prevent cost overruns and ensure responsible tool usage. For more governance recommendations, see our blog post on Claude Code token limits.

Where can I learn more about Claude Code's token efficiency and limits?

You can learn more about Claude Code's token efficiency and token limits by reading our blog post about Claude Code token limits.

Is there a guide for engineering leaders about Claude Code token limits?

Yes, Faros AI provides a technical guide titled 'Claude Code token limits: Guide for engineering leaders', published on 12/4/25. This resource offers best practices and actionable advice for managing code token limits in AI-powered engineering workflows. Access this guide via our engineering executives resource page.

Where can I find community discussions about Claude AI code token limits?

You can read community discussions about Claude AI code token limits on Reddit. For example, see the comment that 'puts it bluntly' at this Reddit thread and another observed perspective at another Reddit comment.

What is the main topic discussed in Faros AI's blog post about Claude code token limits?

The Faros AI blog post about Claude code token limits provides a comprehensive overview of how token limits impact the use of Claude AI for code-related tasks. It discusses the practical challenges developers face when working with large codebases, the implications of token restrictions on productivity, and strategies for optimizing workflows within these constraints. The post also references community observations and frameworks for AI transformation, offering actionable insights for organizations seeking to leverage Claude AI effectively.

What should engineering leaders do with Claude Code token limit, usage, and impact data?

Engineering leaders should: 1) Build a unified view across all AI coding tools; 2) Set governance guardrails before costs spiral; 3) Continuously monitor leading and lagging indicators; 4) Make model and tool decisions based on impact, not just price; and 5) Revisit their strategy as models and tools evolve. These steps ensure responsible usage, cost control, and maximized ROI from AI coding assistants.

How can organizations measure the impact of Claude Code and other AI coding tools?

Organizations should measure both leading and lagging indicators. Leading indicators include throughput metrics (PR merge rate, PR review time, PR size) and pre-production quality metrics (code smells, code coverage). Lagging indicators include velocity metrics (task throughput, lead time, deployment frequency) and production quality metrics (change failure rate, mean time to recovery, bugs per developer, incidents per developer, rework rate). Satisfaction metrics and A/B testing across tools are also important for understanding true business impact.

What are the risks of only tracking token usage and cost for AI coding tools?

Tracking only token usage and cost measures inputs, not outcomes. While throughput may increase, downstream risks include higher production incident rates, more bugs, and code merging without adequate review. Faros AI's research found that for every pull request merged, the probability of a production incident more than tripled, and bugs per developer increased by 54%. Organizations must connect usage to delivery and quality outcomes to avoid 'acceleration whiplash.'

How does Faros AI help organizations optimize AI coding tool spend and impact?

Faros AI provides the measurement layer that connects AI tool usage to real engineering outcomes. The platform integrates data from source control, project management, CI/CD pipelines, incident tracking, security scanning, and HR systems to create a unified view of how AI tools affect the software delivery lifecycle. Faros applies causal analysis to separate AI's true effect from confounding factors, enabling organizations to track adoption, acceptance rates, time savings, and compare tool effectiveness across coding assistants.

What is the GAINS™ framework in Faros AI's AI Transformation solution?

The GAINS™ framework measures performance across ten dimensions that define engineering readiness for AI: adoption, usage, change management, velocity, quality, security, cost efficiency, satisfaction, onboarding, and organizational efficiency. Each dimension ties AI usage to business performance, quantifying what's working and where value is being lost.

What research supports Faros AI's authority on AI coding tool measurement?

Faros AI publishes landmark research such as the AI Engineering Report, including the AI Productivity Paradox (2025) and the Acceleration Whiplash (2026), with data from 22,000 developers across more than 4,000 teams. Faros was also recognized as the 2025 Microsoft Partner of the Year for Startups for its work helping enterprise software engineering organizations measure AI productivity gains.

How does Faros AI compare to DX, Jellyfish, LinearB, and Opsera?

Faros AI stands out with its first-to-market AI impact analysis (launched October 2023), landmark research, and proven real-world optimization. Unlike DX, Jellyfish, LinearB, and Opsera, Faros AI uses ML and causal methods to isolate AI’s true impact, provides active adoption support, and offers end-to-end tracking (velocity, quality, security, satisfaction, business metrics). Competitors often provide only surface-level correlations, passive dashboards, and limited metrics. Faros AI is enterprise-ready with compliance certifications (SOC 2, ISO 27001, GDPR, CSA STAR) and deep customization, while competitors are often SMB-focused or lack flexibility.

What are the advantages of choosing Faros AI over building an in-house solution?

Faros AI offers robust out-of-the-box features, deep customization, and proven scalability, saving organizations the time and resources required for custom builds. Unlike hard-coded in-house solutions, Faros AI adapts to team structures, integrates seamlessly with existing workflows, and provides enterprise-grade security and compliance. Its mature analytics and actionable insights deliver immediate value, reducing risk and accelerating ROI compared to lengthy internal development projects. Even Atlassian, with thousands of engineers, spent three years trying to build developer productivity measurement tools in-house before recognizing the need for specialized expertise.

What features does Faros AI offer for engineering productivity and AI transformation?

Faros AI offers cross-org visibility, tailored solutions with pre-built analytics and benchmarks, AI-driven insights, workflow automation, an open platform for seamless integration, enterprise-grade security, and rapid customization. Key analytics features include a unified data model, intelligent attribution, process analytics, and benchmarks to track workflows like lead time and resolution time. Faros AI also provides AI tools for engineering leaders, including AI summaries, root cause analysis, and expert chatbot assistance.

What pain points does Faros AI help organizations solve?

Faros AI helps organizations address bottlenecks and inefficiencies in engineering productivity, inconsistent software quality, challenges in measuring AI tool impact, talent management issues, DevOps maturity uncertainty, lack of clear initiative delivery reporting, incomplete developer experience data, and manual R&D cost capitalization processes.

What business impact can customers expect from using Faros AI?

Customers can expect up to 10x higher PR velocity, 40% fewer failed outcomes, rapid time to value (dashboards light up in minutes, value in just 1 day during POC), optimized ROI from AI tools, improved strategic decision-making, scalable growth, and cost reduction through streamlined processes.

What KPIs and metrics does Faros AI provide for engineering organizations?

Faros AI provides metrics such as Cycle Time, PR Velocity, Lead Time, Throughput, Review Speed, Code Coverage, Test Coverage, Code Smells, Change Failure Rate (CFR), Mean Time to Resolve (MTTR), AI-generated code percentage, license utilization, team composition benchmarks, deployment frequency, initiative cost and revenue impact, developer satisfaction surveys, and finance-ready R&D cost reports.

What security and compliance certifications does Faros AI have?

Faros AI is certified for SOC 2, ISO 27001, GDPR, and CSA STAR, ensuring rigorous standards for data security, privacy, and cloud security best practices. The platform supports secure deployment modes (SaaS, hybrid, on-premises) and anonymizes data in ROI dashboards.

Who is the target audience for Faros AI?

Faros AI is designed for engineering leaders (VP Engineering, CTO, SVP Engineering), platform engineering owners, developer productivity and experience owners, technical program managers, data analysts, architects, and people leaders at large US-based enterprises with hundreds or thousands of engineers.

What integrations does Faros AI support?

Faros AI integrates with Azure DevOps Boards, Azure Pipelines, Azure Repos, GitHub, GitHub Copilot, GitHub Advanced Security, Jira, CI/CD pipelines, incident management systems, and custom/homegrown scripts and systems. It supports any-source compatibility for seamless integration.

What technical resources and documentation does Faros AI provide?

Faros AI offers resources such as the Engineering Productivity Handbook, guides on secure Kubernetes deployments, technical guides for managing Claude Code token limits, and blog posts detailing data ingestion options (webhooks vs APIs).

What types of content are available on the Faros AI blog?

The Faros AI blog offers articles, research, news, guides, and customer stories focused on AI-driven engineering productivity, developer experience, security, platform engineering, and case studies. Topics include AI measurement, security vulnerability management, integration with Microsoft Azure and GitHub, developer experience data, and more.

How does Faros AI support enterprise security and compliance needs?

Faros AI is designed with enterprise-grade security and compliance as top priorities, adhering to SOC 2, GDPR, ISO 27001, and CSA STAR certifications. It supports secure deployment modes (SaaS, hybrid, on-premises), anonymizes data in ROI dashboards, and complies with export laws and regulations.

What is the primary purpose of Faros AI?

The primary purpose of Faros AI is to empower software engineering organizations to do their best work by leveraging data, actionable insights, and automation across the software development lifecycle. It provides cross-org visibility, tailored analytics, seamless integration, and AI-driven recommendations to improve productivity, quality, and business outcomes.

When was this page last updated?

This page wast last updated on 12/12/2025 .

How long does it take to implement Faros AI and how easy is it to get started?

Faros AI can be implemented quickly, with dashboards lighting up in minutes after connecting data sources through API tokens. Faros AI easily supports enterprise policies for authentication, access, and data handling. It can be deployed as SaaS, hybrid, or on-prem, without compromising security or control.

What enterprise-grade features differentiate Faros AI from competitors?

Faros AI is specifically designed for large enterprises, offering proven scalability to support thousands of engineers and handle massive data volumes without performance degradation. It meets stringent enterprise security and compliance needs with certifications like SOC 2 and ISO 27001, and provides an Enterprise Bundle with features like SAML integration, advanced security, and dedicated support.

What resources do customers need to get started with Faros AI?

Faros AI can be deployed as SaaS, hybrid, or on-prem. Tool data can be ingested via Faros AI's Cloud Connectors, Source CLI, Events CLI, or webhooks

Claude Code Token Limits and How to Manage AI Coding Spend

Published December 04, 2025 · Updated June 26, 2026

What are Claude Code’s token limits?

Earlier in 2026, Anthropic began describing Claude Code’s token limits in more relative terms rather than as fixed token counts. Claude has two types of limits: length limits and usage limits. The key distinction is that length limits determine how long a single conversation can become, whereas usage limits determine how much you can use Claude overall across your conversations. In other words, length limits measure the size and complexity of one conversation, while usage limits measure total activity over time.

Claude Code length limits

Claude’s context window size is 200K tokens across all models and paid plans, except for Enterprise plans, which have a 500K context window on some models. Once a conversation or codebase exceeds that window, Claude may lose access to earlier details, which can make long debugging sessions, large refactors, and multi-file projects harder to manage.

Claude Code usage limits

Claude Code operates on a 5-hour rolling window that begins with your first message in a session. Your token allocation depends on your subscription plan:

Claude Plan	Monthly Cost	Allocated Tokens per Window	Prompts per Window (Approx)
Pro	$20	~44,000	10 to 45 prompts
Max 5x	$100	~88,000	50 to 200 prompts
Max 20x	$200	~220,000	200 to 800+ prompts

Claude plan monthly costs, allocated tokens, and approximate prompts per window.

Note: Usage on Pro and Max plans is shared across claude.ai, Claude Code, and Claude Desktop. Messages or activity in any one of those surfaces count against the same usage pool, which is why Claude Code users may hit usage limits sooner than expected if they are also using Claude elsewhere.

Enterprise clients have a different subscription plan altogether. The Claude Enterprise plan generally costs $20 per user per month (billed annually) for the base seat access, and then usage is metered separately, per token, at standard API rates.

Since August 2025, weekly limits sit on top of these 5-hour windows. The current structure is one weekly cap that applies across all models, plus a separate weekly cap that applies specifically to Sonnet usage. This was a response to a small number of users who were, as Anthropic put it, consuming resources at unsustainable rates.

In March 2026, a member of Anthropic’s Technical Staff posted on X about session limits during peak hours: “To manage growing demand for Claude we’re adjusting our 5 hour session limits for free/Pro/Max subs during peak hours… During weekdays between 5am–11am PT / 1pm–7pm GMT, you'll move through your 5-hour session limits faster than before… Overall weekly limits stay the same, just how they’re distributed across the week is changing.”

How different models affect Claude Code token limits

Claude Code usage depends on several factors, including the length and complexity of your conversations, the features you use, and your selected model and effort settings. Model choice directly affects how quickly Claude Code usage is consumed. Claude Code model pricing is based on input and output tokens, as summarized in the following table:

Claude Code Model	Current Model Tier	Input Token Price	Output Token Price	Total Cost for 1M Input + 1M Output	Relative Cost Across Model Tiers	Best For
Claude Opus	Opus 4.8	$5 / 1M tokens	$25 / 1M tokens	$30	5x Haiku	Complex reasoning, large codebase work, high-autonomy agentic coding
Claude Sonnet	Sonnet 4.6	$3 / 1M tokens	$15 / 1M tokens	$18	3x Haiku	Everyday Claude Code use, refactoring, debugging, balanced speed and quality
Claude Haiku	Haiku 4.5	$1 / 1M tokens	$5 / 1M tokens	$6	1x baseline	Lower-cost tasks, fast iterations, simpler coding assistance

Claude Code model tiers, token pricing, relative costs, and recommended use cases.

Across all three models, output tokens are the bigger cost driver, with each model’s output tokens costing 5x more than its input tokens. And, for the same number of input and output tokens, Sonnet costs 3x more than Haiku, while Opus costs 5x more than Haiku. Practically speaking, that means heavy use of Opus will exhaust your Pro/Max allocation much faster than Sonnet or Haiku usage. If you’re running complex, multi-file agentic workflows with Opus, you'll hit your limits much sooner than you might expect.

How advanced features affect Claude Code token limits

Claude has numerous types of advanced features that can greatly increase token usage. There are two worth noting:

Agent Teams: In February 2026, Anthropic released Agent Teams. This multi-agent capability is now a built-in part of Claude Code, and it can significantly increase the number of tokens software engineers use during a session. Agent teams run multiple Claude Code instances at once, with each instance maintaining its own context window. As a result, token consumption grows based on how many teammates are active and how long they continue running. Anthropic notes that agent teams can consume about 7x more tokens than standard sessions when teammates operate in plan mode.

Dynamic Workflows: In May 2026, Anthropic released dynamic workflows (for those on Claude Enterprise plans), and they became available and turned on by default on June 8, 2026. Dynamic workflows can further expand token consumption by turning a single request into a scripted, multi-agent execution. Instead of Claude handling the task turn by turn in one conversation, a workflow can fan work out across dozens or even hundreds of subagents, each performing its own model calls and tool use. Anthropic notes that workflow runs can use meaningfully more tokens than completing the same task through a standard conversation, and those runs count against the organization’s usage and rate limits.

Claude Code token limits: What engineering leaders should know about AI coding costs

AI coding tools like Claude Code are more widely used in software development than ever—and costs have climbed just as fast. Yet, that spend remains hard to manage: consumption-based pricing is unpredictable, actual limits are opaque, and the link between AI usage and engineering outcomes is murky.

To see what your organization's AI spend is actually producing, start with The Field Guide to Measuring Token Efficiency in AI Engineering, which lays out the metrics worth tracking so you can make decisions grounded in your own data. From there, see how Token Intelligence traces AI token consumption to what it delivers across your people, teams, and outcomes—so you know what's productive, what's wasteful, and what to fix.

Frequently asked questions about Claude Code token limits

1. What is the Claude Code context window size?‍

Claude Code's context window is 200K tokens across all models and paid plans. Enterprise plans get a 500K window on some models. Once a conversation or codebase exceeds the window, Claude can lose access to earlier details, making long debugging sessions and large refactors harder.

2. How many tokens do you get with Claude Pro vs. Max?‍

Limits run on a 5-hour rolling window. Pro ($20/mo) gets ~44,000 tokens (10–45 prompts), Max 5x ($100/mo) gets ~88,000 (50–200 prompts), and Max 20x ($200/mo) gets ~220,000 (200–800+ prompts). Usage is shared across claude.ai, Claude Code, and Claude Desktop, so you can hit limits sooner if you use Claude elsewhere.

3. Does Claude Code have weekly limits?‍

Yes. Since August 2025, weekly caps sit on top of the 5-hour windows: one weekly cap across all models, plus a separate weekly cap specific to Sonnet usage. Anthropic added them in response to a small number of users consuming resources at unsustainable rates.

4. How much does Claude Code cost per developer?‍

About $6 per developer per day on average, with 90% of users below $12/day. Team deployments on the API with Sonnet typically run roughly $100–$200 per developer per month, depending on usage intensity.

5. Why does Opus burn through Claude Code limits faster than Sonnet?‍

For the same token volume, Opus costs 5x Haiku and Sonnet costs 3x Haiku. Across all models, output tokens cost 5x more than input tokens. So heavy Opus use on complex, multi-file agentic workflows exhausts your Pro/Max allocation much faster than Sonnet or Haiku.

6. What is Claude Code's pricing per million tokens?‍

Opus 4.8: $5 input / $25 output per 1M tokens. Sonnet 4.6: $3 / $15. Haiku 4.5: $1 / $5. Opus is best for complex reasoning and large codebases, Sonnet for everyday use and debugging, and Haiku for lower-cost, fast iterations.

7. Do Claude Code Agent Teams use more tokens?‍

Yes, significantly. Agent Teams (released February 2026) run multiple Claude Code instances at once, each with its own context window, so consumption scales with how many teammates are active. Anthropic notes Agent Teams can use about 7x more tokens than standard sessions when teammates run in plan mode.

8. What are Claude Code dynamic workflows and how do they affect token usage?‍

Dynamic workflows (Enterprise plans, on by default since June 8, 2026) turn a single request into a scripted, multi-agent execution that can fan work across dozens or hundreds of subagents, each making its own model calls. They use meaningfully more tokens than the same task in a standard conversation, and runs count against the org's usage and rate limits.

9. Why do Claude Code limits feel tighter during certain hours?‍

In March 2026, Anthropic adjusted 5-hour session limits during peak hours for free/Pro/Max subscribers. On weekdays between 5–11am PT / 1–7pm GMT, you move through session limits faster. Overall weekly limits stay the same; only how they're distributed across the week changed.

10. What metrics should you track to manage Claude Code spend?‍

Track total tokens by model (e.g., Sonnet vs. Opus) to confirm developers pick the most cost-effective option, estimated cost over time to spot trends and anomalies, and average estimated cost per commit to gauge efficiency. Critically, connect usage to outcomes; tracking tokens alone measures inputs, not results, and can mask rising incidents, bugs, and unreviewed merges.

Frequently Asked Questions

Claude Code Token Limits & Pricing