Claude guide
Model prices, task choices, and plans. Compare Anthropic's API costs and Claude subscriptions in one reference.
Reviewed against Anthropic's documentation on . Prices are in USD, before tax.
What does each model cost?
Anthropic pricing documentationThese are direct Claude API token rates, not the price of a Claude chat message or a subscription allowance. Start with the current lineup, then select older public models or restricted models when you need them.
Prices use standard global routing. Search by model name or API ID, and compare Standard, Batch, or Fast mode. Models listed here can have different availability and prices in GitHub Copilot or cloud-provider catalogs.
Scroll to compare all columns. Model names and API IDs stay in view.
| Model / API ID | Mode / prompt length | Input | Cache read | 5-minute write | 1-hour write | Output |
|---|---|---|---|---|---|---|
Claude Haiku 5.5 claude-haiku-5-5 Current lineup · active · See note | Batch APIPrompt ≤ 100K tokens | $0.05 | $0.005 | $0.0625 | $0.10 | $0.25 |
Claude Haiku 5.5 claude-haiku-5-5 Current lineup · active · See note | Standard APIPrompt ≤ 100K tokens | $0.10 | $0.01 | $0.125 | $0.20 | $0.50 |
Claude Haiku 5.5 claude-haiku-5-5 Current lineup · active · See note | Batch APIPrompt > 100K tokens | $0.25 | $0.025 | $0.3125 | $0.50 | $1.25 |
Claude Haiku 4.5 claude-haiku-4-5-20251001 Earlier version · active | Batch APIStandard prompt pricing | $0.50 | $0.05 | $0.625 | $1.00 | $2.50 |
Claude Haiku 5.5 claude-haiku-5-5 Current lineup · active · See note | Standard APIPrompt > 100K tokens | $0.50 | $0.05 | $0.625 | $1.00 | $2.50 |
Claude Haiku 4.5 claude-haiku-4-5-20251001 Earlier version · active | Standard APIStandard prompt pricing | $1.00 | $0.10 | $1.25 | $2.00 | $5.00 |
Claude Sonnet 5 claude-sonnet-5 Earlier version · active | Batch APIStandard prompt pricing | $1.00 | $0.10 | $1.25 | $2.00 | $5.00 |
Claude Sonnet 5.5 claude-sonnet-5-5 Current lineup · active | Batch APIStandard prompt pricing | $1.00 | $0.05 | $1.25 | $2.00 | $5.00 |
Claude Sonnet 4.5 claude-sonnet-4-5-20250929 Deprecated · retires 30 Nov 2026 · See note | Batch APIStandard prompt pricing | $1.50 | $0.15 | $1.875 | $3.00 | $7.50 |
Claude Sonnet 4.6 claude-sonnet-4-6 Earlier version · active | Batch APIStandard prompt pricing | $1.50 | $0.15 | $1.875 | $3.00 | $7.50 |
Claude Opus 5.5 claude-opus-5-5 Current lineup · active | Batch APIStandard prompt pricing | $2.00 | $0.10 | $2.50 | $4.00 | $10.00 |
Claude Sonnet 5 claude-sonnet-5 Earlier version · active | Standard APIStandard prompt pricing | $2.00 | $0.20 | $2.50 | $4.00 | $10.00 |
Claude Sonnet 5.5 claude-sonnet-5-5 Current lineup · active | Standard APIStandard prompt pricing | $2.00 | $0.10 | $2.50 | $4.00 | $10.00 |
Claude Opus 4.5 claude-opus-4-5-20251101 Earlier version · active | Batch APIStandard prompt pricing | $2.50 | $0.25 | $3.125 | $5.00 | $12.50 |
Claude Opus 4.6 claude-opus-4-6 Earlier version · active | Batch APIStandard prompt pricing | $2.50 | $0.25 | $3.125 | $5.00 | $12.50 |
Claude Opus 4.7 claude-opus-4-7 Earlier version · active | Batch APIStandard prompt pricing | $2.50 | $0.25 | $3.125 | $5.00 | $12.50 |
Claude Opus 4.8 claude-opus-4-8 Earlier version · active | Batch APIStandard prompt pricing | $2.50 | $0.25 | $3.125 | $5.00 | $12.50 |
Claude Opus 5 claude-opus-5 Earlier version · active | Batch APIStandard prompt pricing | $2.50 | $0.25 | $3.125 | $5.00 | $12.50 |
Claude Sonnet 4.5 claude-sonnet-4-5-20250929 Deprecated · retires 30 Nov 2026 · See note | Standard APIStandard prompt pricing | $3.00 | $0.30 | $3.75 | $6.00 | $15.00 |
Claude Sonnet 4.6 claude-sonnet-4-6 Earlier version · active | Standard APIStandard prompt pricing | $3.00 | $0.30 | $3.75 | $6.00 | $15.00 |
Claude Opus 5.5 claude-opus-5-5 Current lineup · active | Standard APIStandard prompt pricing | $4.00 | $0.20 | $5.00 | $8.00 | $20.00 |
Claude Fable 5 claude-fable-5 Earlier version · active · See note | Batch APIStandard prompt pricing | $5.00 | $0.50 | $6.25 | $10.00 | $25.00 |
Claude Fable 5.1 claude-fable-5-1 Current lineup · active · See note | Batch APIStandard prompt pricing | $5.00 | $0.125 | $6.25 | $10.00 | $25.00 |
Claude Mythos 5 claude-mythos-5 Verified organizations only · See note | Batch APIStandard prompt pricing | $5.00 | $0.50 | $6.25 | $10.00 | $25.00 |
Claude Mythos 5.1 claude-mythos-5-1 Verified organizations only · See note | Batch APIStandard prompt pricing | $5.00 | $0.125 | $6.25 | $10.00 | $25.00 |
Claude Opus 4.5 claude-opus-4-5-20251101 Earlier version · active | Standard APIStandard prompt pricing | $5.00 | $0.50 | $6.25 | $10.00 | $25.00 |
Claude Opus 4.6 claude-opus-4-6 Earlier version · active | Standard APIStandard prompt pricing | $5.00 | $0.50 | $6.25 | $10.00 | $25.00 |
Claude Opus 4.7 claude-opus-4-7 Earlier version · active | Standard APIStandard prompt pricing | $5.00 | $0.50 | $6.25 | $10.00 | $25.00 |
Claude Opus 4.8 claude-opus-4-8 Earlier version · active | Standard APIStandard prompt pricing | $5.00 | $0.50 | $6.25 | $10.00 | $25.00 |
Claude Opus 5 claude-opus-5 Earlier version · active | Standard APIStandard prompt pricing | $5.00 | $0.50 | $6.25 | $10.00 | $25.00 |
Claude Opus 5.5 claude-opus-5-5 Current lineup · active | Fast modeFull context window | $8.00 | $0.40 | $10.00 | $16.00 | $40.00 |
Claude Fable 5 claude-fable-5 Earlier version · active · See note | Standard APIStandard prompt pricing | $10.00 | $1.00 | $12.50 | $20.00 | $50.00 |
Claude Fable 5.1 claude-fable-5-1 Current lineup · active · See note | Standard APIStandard prompt pricing | $10.00 | $0.25 | $12.50 | $20.00 | $50.00 |
Claude Mythos 5 claude-mythos-5 Verified organizations only · See note | Standard APIStandard prompt pricing | $10.00 | $1.00 | $12.50 | $20.00 | $50.00 |
Claude Mythos 5.1 claude-mythos-5-1 Verified organizations only · See note | Standard APIStandard prompt pricing | $10.00 | $0.25 | $12.50 | $20.00 | $50.00 |
Claude Opus 4.8 claude-opus-4-8 Earlier version · active | Fast modeFull context window | $10.00 | $1.00 | $12.50 | $20.00 | $50.00 |
Claude Opus 5 claude-opus-5 Earlier version · active | Fast modeFull context window | $10.00 | $1.00 | $12.50 | $20.00 | $50.00 |
No matching rates. Try another model, change the model range or pricing mode, or reset the filters.
Batch cache rates and fast-mode cache rates are calculated from Anthropic's published discount and caching rules. Fast mode cannot be combined with Batch. A cache write or read replaces the base input charge for those tokens; these columns are not fees to add to every input token.
- Haiku 5.5: prompts above 100,000 input tokens use the higher rate for the request. The threshold counts all input, including cache reads and writes.
- Fable and Mythos: these Covered Models require 30-day data retention unless Anthropic expressly authorizes an exception. Read the retention and access requirements before selecting them.
- Mythos: access requires an organization verified through Anthropic's verification programs. A published price does not grant access; the retention requirement above also applies.
- Sonnet 4.5: deprecated and scheduled to retire on 30 November 2026. It appears only under “All listed models.” Retired models are excluded from this guide.
Sources: token and cache pricing, current model specifications, and model lifecycle.
What can change the bill?
| Factor | How to read it |
|---|---|
| Prompt caching | A 5-minute write is 1.25× base input; a 1-hour write is 2×. Read discounts vary by model, so compare the explicit cache columns. |
| Batch API | Asynchronous processing receives a 50% token-price discount. Prompt-caching discounts stack; Batch is separate from Fast mode. |
| Fast mode | Opus 5.5, Opus 5, and Opus 4.8 have premium rates. Opus 5 and 4.8 fast mode remain in research preview. Eligibility and dedicated limits apply. |
| Long prompts | For 4.6-and-later models, full context uses standard per-token prices except Haiku 5.5's higher band. A larger prompt still costs more because it contains more tokens. |
| US-only inference | Supported 4.6-and-later models use a 1.1× multiplier across token categories. The table shows global rates. |
| Thinking and tools | Reasoning and tool interactions consume tokens; server tools and managed-agent infrastructure can add charges. Token prices alone are not a complete agent-session estimate. |
Rules and exceptions: Anthropic's feature-specific pricing.
Choose for the task
Anthropic model guidanceAnthropic recommends Opus 5.5 as a starting point for most workloads. An efficiency-first approach can start with Haiku; Sonnet balances speed and capability. Evaluate Fable for demanding work when Opus at higher effort still falls short.
- Haiku 5.5
- Extraction, routing, and fast responses
- Sonnet 5.5
- Everyday coding, writing, and analysis
- Opus 5.5
- Complex coding and sustained agent work
- Fable 5.1
- Demanding reasoning and long-horizon tasks
| Your task | Haiku 5.5 | Sonnet 5.5 | Opus 5.5 | Fable 5.1 |
|---|---|---|---|---|
| High-volume classification or extraction | Start here | Also fits | Not first choice | Not first choice |
| Quick syntax help and small edits | Start here | Also fits | Not first choice | Not first choice |
| Everyday features, tests, and documentation | Also fits | Start here | Also fits | Not first choice |
| Writing, analysis, and visual understanding | Also fits | Start here | Also fits | Also fits |
| Difficult debugging across a codebase | Not first choice | Also fits | Start here | Also fits |
| Large refactors and autonomous coding | Not first choice | Also fits | Start here | Also fits |
| Demanding reasoning after Opus evaluation falls short | Not first choice | Not first choice | Also fits | Start here |
| Deep research carried through to a finished artifact | Not first choice | Also fits | Also fits | Start here |
This is not a benchmark ranking. The four current API models accept text and images, with a 1M-token context window and up to 128K output tokens. Those API maxima do not define the allowance in a Claude app plan. Adjust effort and measure quality, latency, and total cost on your own workload.
Sources: selection and evaluation, capabilities and limits.
Which Claude plan fits?
Claude plan documentationChoose a subscription for using Claude and Claude Code, or API billing for an application you build. Eligible Max and Team plans also offer separately claimed API credits. These balances and usage limits are different.
Personal plans
| Compare | Free | Pro | Max 5x | Max 20x |
|---|---|---|---|---|
| Monthly price (USD) | $0 | $20 | $100 | $200 |
| Annual billing | Not applicable | $200 / year, paid upfront | Monthly billing only | Monthly billing only |
| Included app usage | Limited | More than Free; session and weekly limits | 5× Pro session capacity; weekly limits still apply | 20× Pro session capacity; weekly limits still apply |
| Model family access | Sonnet and Haiku | Adds Opus | Opus, Sonnet, and Haiku | Opus, Sonnet, and Haiku |
| Fable access | Not included | Paid usage credits | Separate limit: 50% of weekly capacity | Separate limit: 50% of weekly capacity |
| Claude Code | Not included | Included; shared usage limits | Included; shared usage limits | Included; shared usage limits |
| Separate monthly API credits | Not included | Not included | $100 / month after claim | $200 / month after claim |
Pro and Max have shared session and weekly limits across Claude and Claude Code. Fable has separate access rules and limits. Once included usage is exhausted, paid usage credits or another billing route can add cost; check the account's current settings.
Team and Enterprise
| Compare | Team Standard | Team Premium | Enterprise |
|---|---|---|---|
| Monthly billing (USD) | $25 / seat | $125 / seat | Annual seat contract + usage |
| Annual price (USD) | $20 / seat / month | $100 / seat / month | $20 / seat / month + usage |
| Seat requirements | 2 seats; teams up to 150 | 2 seats; can mix Team seat types | 20 seats on published self-serve offer |
| Usage model | More usage than Pro; limits apply | 5× Standard seat usage; limits apply | Metered usage at API rates |
| Claude Code | Included | Included | Included; usage billed separately |
| Fable access | Check seat eligibility; pricing lists the allowance for Premium seats | Separate limit: 50% of weekly capacity | Available, subject to organization controls |
| Monthly API credit contribution | $20 / seat, pooled and capped | $100 / seat, pooled and capped | Not included in this credit program |
| Organization controls | Central billing, SSO, organization administration | Same Team administration with higher seat usage | Spend controls, SCIM, audit logs, custom retention |
The Enterprise price above is the published self-serve offer. Sales-assisted terms can differ. Team API credits are capped at $500 per month for the whole plan, not $500 per seat.
Claim API credits separately. Eligible new Max and Team subscribers can claim after seven days by linking a Console organization. Credits expire each cycle without rollover. Free, Pro, and Enterprise do not qualify for this program.
Keep the billing routes clear. Included API credits cover Claude Platform applications and agents. They do not cover Claude Code, extra usage in Claude apps, or partner clouds. API-key authentication in Claude Code uses API billing instead of subscription usage.
Training is separate from retention. Commercial inputs and outputs are not used for training by default; voluntary feedback and permission are exceptions. Personal accounts have model-improvement choices, with feedback and safety-review exceptions. Fable and Mythos retention still needs a separate check.
Details: personal tier prices, API-credit rules, Claude Code limits and billing, commercial data use, and consumer data use.
Sources & updates
This reference is manually maintained, last reviewed on . Model availability, promotional benefits, usage limits, and prices can change. Check the linked official sources before choosing a plan or setting a budget.
- API token prices, caching, and pricing modifiers
- Current models, API IDs, and capabilities
- Model lifecycle and retirement dates
- Anthropic model selection guidance
- Claude app and organization pricing
- Individual plans and Max tier prices
- Included API credits: eligibility and restrictions
- Claude Code subscription limits and API billing
- Fable and Mythos access and retention requirements
- Commercial data use and model training
- Consumer model-improvement choices and exceptions
The task matrix is an editorial synthesis of Anthropic's guidance. This page is independent of Anthropic and is not official documentation. For Claude models billed through Copilot, use the GitHub Copilot guide.
Back to all resources