Claude Code Pricing
Complete Claude AI price breakdown including Pro (from $17/month annually) vs Max (from $100/month) subscription plans, API costs, and usage limits. Choose the right Claude pricing plan for your development workflow and budget.
Claude AI Price Comparison: Pro vs Max Plans
| Plan | Price | Claude Models | Usage Limits | Best For | Claude Code Rating |
|---|---|---|---|---|---|
| Free | $0/month | Limited Sonnet access | Very limited | Multi-platform access, web search | |
| Pro | $17/month annually ($200 upfront) $20/month monthly | Sonnet 5 + multiple models | 5x free tier | Development + Research + Integrations | |
| Max $100 | $100/month | Sonnet 5 + Opus 5 + Fable 5 | 5x Pro limits + priority access | Professional + early features | |
| Max $200 | $200/month | All models + full Fable 5 | 20x Pro limits + priority access | Full professional + early features | |
| API | Pay-per-use | All models | Usage-based billing | Custom needs |
Detailed Plan Breakdown
Claude Free - Limited Testing
The free plan provides limited Sonnet access for basic usage only, with very low usage limits that are quickly exhausted. It includes multi-platform access (web, iOS, Android, desktop), web search capability, and desktop extensions. However, the free plan does not support Claude Code access. You need at least a Pro subscription or API credits to use Claude Code.
Claude Pro ($17/month annually, $20/month monthly) - Development + Research + Integrations
Claude Pro includes access to Claude Sonnet 5 model and additional models with 5x usage limits compared to the free tier. It includes Research access, Google Workspace integration (email, calendar, docs), remote MCP server connections for tool integrations, and extended thinking for complex work. However, it lacks included access to Claude Opus 5 and Fable 5 (Fable 5 is available on Pro via metered usage credits). For Claude Code, this plan works well for development, small projects, and learning, though you may hit usage limits with extensive coding sessions. Annual billing ($200 upfront) saves $36 per year compared to monthly billing.
Best For:
- Learning Claude Code - Affordable way to get started
- Hobby projects - Small personal development work with research needs
- Workspace integration - Teams using Google Workspace
- Tool connectivity - Developers wanting MCP server connections
- Budget constraints - Cost-effective entry point with full features
Claude Max $100 ($100/month) - Professional Development with Efficient Opus Access
Claude Max $100 provides access to Claude Sonnet 5 with 5x higher usage limits than Pro, plus access to Claude Fable 5 (state-of-the-art FrontierCode and vision, 50M-line Ruby migrations in a day) and the current Opus tier, Claude Opus 5 (doubles Opus 4.8 on Frontier-Bench v0.1, mid-conversation tool changes, agentic self-verification, same $5/$25 pricing as Opus 4.8). It includes higher output limits for all tasks, early access to advanced Claude features, and priority access during high traffic times. As of July 20, 2026, Fable 5 is included on Max and Team Premium plans at up to 50% of weekly usage limits, while Pro and Team Standard access Fable 5 via metered usage credits (a one-time $100 promotional credit is claimable July 20 through August 2, expiring September 17, 2026).
Best For:
- Professional developers - Regular Claude Code usage with occasional complex needs
- Medium projects - Projects that need some advanced analysis but not constantly
- Budget-conscious professionals - Want Opus access without full Max cost
- Priority access needs - Developers who need reliable access during peak times
Claude Max $200 ($200/month) - Full Professional Development
Claude Max provides access to all Claude models including Claude Fable 5, with 20x higher usage limits than Pro. Fable 5 leads Cognition's FrontierCode benchmark, compressed a 50-million-line Ruby migration into a single day (Stripe), achieves state-of-the-art vision (including vision-only Pokémon FireRed completion), and shows 3× stronger long-context memory improvement than Opus 4.8. Claude Opus 5, the current Opus tier, is also available, doubling Opus 4.8's score on Frontier-Bench v0.1, landing within 0.5% of Fable 5 on CursorBench 3.2 at half the cost, and adding mid-conversation tool changes and agentic self-verification at the same $5/$25 pricing as Opus 4.8. Opus 4.8 remains available with ~4× fewer overlooked code flaws, Dynamic Workflows (hundreds of parallel subagents in Claude Code), effort control (standard/extra/max), mid-task cache preservation via the Messages API, and fast mode at 3× lower cost than Opus 4.7. It includes higher output limits for all tasks, early access to advanced Claude features, and priority access during high traffic times. For Claude Code, this plan handles professional development with large projects and complex tasks, extended coding sessions without hitting limits, and full access to Fable 5 for frontier capability.
Best For:
- Professional developers - Daily Claude Code usage who want the best model
- Large projects - Complex codebases and architecture work
- Maximum productivity - Unrestricted Claude Code workflow
- Early adopters - Developers who want first access to new features
API Pricing - Pay-Per-Use
Current API Rates (July 2026)
| Model | Input Tokens | Output Tokens | Context Window | Best For |
|---|---|---|---|---|
| Claude Fable 5 | $10.00 / 1M | $50.00 / 1M | — | Latest frontier model, state-of-the-art FrontierCode and vision, codebase-scale migrations, million-token long-context focus |
| Claude Opus 5 | $5.00 / 1M ($10.00 / 1M fast mode) | $25.00 / 1M ($50.00 / 1M fast mode) | — | Current Opus tier, doubles Opus 4.8 on Frontier-Bench v0.1, mid-conversation tool changes, agentic self-verification |
| Claude Opus 4.8 | $5.00 / 1M ($10.00 / 1M fast mode) | $25.00 / 1M ($50.00 / 1M fast mode) | — | Previous Opus generation, ~4× fewer flaw misses, dynamic workflows, effort control |
| Claude Opus 4.7 | $5.00 / 1M | $25.00 / 1M | — | Previous Opus generation, 3× vision, xhigh effort |
| Claude Opus 4.5 | $5.00 / 1M | $25.00 / 1M | 200K tokens | Legacy, state-of-the-art SWE-bench, agents |
| Claude Sonnet 5 | $3.00 / 1M ($2.00 / 1M through Aug 31, 2026) | $15.00 / 1M ($10.00 / 1M through Aug 31, 2026) | 1M tokens (native) | Latest daily-driver coding model, approaches Opus 4.8 on agentic/tool-use benchmarks |
| Claude Sonnet 4.6 | $3.00 / 1M ($6.00 / 1M for prompts >200K) | $15.00 / 1M ($22.50 / 1M for prompts >200K) | 1M tokens* | Previous Sonnet generation, large codebases |
| Claude Sonnet 4.5 | $3.00 / 1M | $15.00 / 1M | 200K tokens | Legacy, cost-effective development |
| Claude Haiku 4.5 | $1.00 / 1M | $5.00 / 1M | 200K tokens | Fast tasks, high volume |
| Claude Haiku 3.5 | $0.80 / 1M | $4.00 / 1M | 200K tokens | Retired, simple tasks |
*Claude Sonnet 4.6's 1M token context window is currently available via API only. This feature is likely coming to Claude Max subscriptions in the future.
Sonnet 5 Pricing: Claude Sonnet 5 launches at $2.00 input / $10.00 output per 1M tokens through August 31, 2026, then reverts to the standard rate of $3.00 input / $15.00 output per 1M tokens — the same standard rate as Sonnet 4.6, despite substantial agentic and coding improvements.
API Benefits: The API provides precise cost control by charging only for actual usage, with access to all models including latest releases and no usage limits for scaling as needed. Sonnet 5 ships with a native 1M token context window (no separate beta flag needed), and Sonnet 4.6 via API retains its own 1M token context window. Enterprise features and custom configurations are also available.
API Considerations: However, costs can be variable and expensive with heavy usage, requiring monitoring of token usage and billing complexity.
Batch API Pricing (50% Discount)
For bulk operations, the Batch API offers 50% savings on all models:
| Model | Batch Input | Batch Output |
|---|---|---|
| Claude Fable 5 | $5.00 / 1M | $25.00 / 1M |
| Claude Opus 5 | $2.50 / 1M | $12.50 / 1M |
| Claude Opus 4.8 | $2.50 / 1M | $12.50 / 1M |
| Claude Opus 4.7 | $2.50 / 1M | $12.50 / 1M |
| Claude Opus 4.5 | $2.50 / 1M | $12.50 / 1M |
| Claude Sonnet 5 | $1.50 / 1M | $7.50 / 1M |
| Claude Sonnet 4.6 | $1.50 / 1M | $7.50 / 1M |
| Claude Haiku 4.5 | $0.50 / 1M | $2.50 / 1M |
Useful for batch code analysis, bulk file processing, and non-real-time development workflows.
Prompt Caching (Up to 90% Savings)
Cache repeated system prompts and CLAUDE.md context to reduce costs:
| Cache Type | Write Cost | Read Cost |
|---|---|---|
| 5-minute cache | 1.25x base input | 0.1x base input |
| 1-hour cache | 2x base input | 0.1x base input |
Prompt caching is especially valuable for Claude Code sessions where system prompts and context are reused across multiple turns.
Tool-Specific API Pricing
| Tool | Additional Cost |
|---|---|
| Web Search | $10 per 1,000 searches |
| Web Fetch | Free (token costs only) |
| Code Execution | 1,550 free hours/month, then $0.05/hour |
| Bash/Text Editor | Standard token costs |
Usage Monitoring and Optimization
Context Management
The new /context command (added in v1.0.86) helps users debug context issues and optimize token usage for better cost management. This command is particularly useful for understanding when you're approaching context limits that could impact pricing.
Track Your Usage
For Subscription Users:
- CC Usage Tool: Install ccusage for detailed tracking
- Monitor patterns: Understand your typical monthly usage
- Optimize workflow: Use appropriate models for different tasks
For API Users:
- Anthropic Console: Monitor usage at console.anthropic.com
- Set alerts: Configure billing alerts to avoid surprises
- Token optimization: Use efficient prompting to reduce costs
Cost Optimization Tips
Model Selection: Use Sonnet 5 for daily development (best capability/cost balance). Switch to Fable 5 for frontier engineering, codebase-scale migrations, vision-heavy work, and long-context agents. Use Opus 5, the current Opus tier, for cost-sensitive Opus-tier reasoning at half the per-token price of Fable 5 (same pricing as Opus 4.8, with roughly double the Frontier-Bench v0.1 score). Use Haiku 4.5 for fast, simple tasks at 1/3 Sonnet's cost. See tactical model selection for detailed guidance.
Batch Processing: For bulk code analysis or non-urgent refactoring, use the Batch API for 50% savings on all models.
Prompt Caching: CLAUDE.md and system prompts are automatically cached. After the first request, cached content costs only 10% of the original price.
Context Management: Use /compact to manage context costs and the /model command to switch models based on task complexity.
For API users:
- Use Sonnet 5 ($3/1M) for routine tasks - best price-to-performance ratio
- Use Fable 5 ($10/1M) for frontier engineering, vision-heavy work, and long-context agents
- Use Opus 5 ($5/1M), the current Opus tier, for complex reasoning at half the per-token cost of Fable 5
- Use Haiku for simple, repetitive operations
For Max subscribers: With current Opus limits (Max users have roughly the same token allocation as previously had with Sonnet), use Opus freely for:
- Complex software engineering tasks requiring state-of-the-art capabilities
- Extended agentic workflows and multi-file reasoning
- Computer use and browser automation
- Any task requiring maximum capability - you won't run out during normal daily usage
See Also: Model Comparison|Installation Guide|Getting Started|Is Claude Code Free|CC Usage Add-on
Updates on Claude Code insights, news, mechanics, features & best practices