Anthropic Claude Pricing in 2026: What You Actually Get on Each Tier
Anthropic operates a seven-tier consumer and business pricing structure across claude.ai plus a per-token API priced separately. As of May 2026 the picture is: Free at $0, Pro at $20/month, Max 5x at $100/month, Max 20x at $200/month, Team Standard at $25/seat/month, Team Premium at $125/seat/month, and Enterprise at custom pricing. The most consequential pricing event in Claude’s API history was the 67% Opus price reduction at the Opus 4.6 launch (2026-02-05): from $15/$75 per million input/output tokens (Opus 4.1) down to $5/$25. The 1M token context window also became standard at no surcharge. Opus 4.7 (2026-04-16) maintained the new $5/$25 pricing.
Two volatile elements are worth flagging up front. First, exact message-volume caps per tier are not publicly published – Anthropic states “usage limits apply” and describes them as a “conversation budget” without specific counts. Second, Claude Code inclusion in the $20 Pro tier is contested: anthropic.com/pricing currently lists “Includes Claude Code” under Pro, while an independent changelog tracker (scriptbyai.com, April 2026) states Anthropic removed Claude Code from Pro. Verify directly at anthropic.com/pricing before making decisions on Pro for Claude Code use.
See also: Claude AI complete guide →
Claude Free Tier: What’s Included
The Free tier costs $0/month and provides access to Claude Sonnet 4.6 (the default model since February 2026) and limited Claude Haiku 4.5. Usage caps are described as “conversation budget” without specific numbers – Anthropic does not publish messages-per-period limits for any tier. Free includes basic web search (toggleable), basic file uploads, Artifacts, and chat memory (memory was extended to Free on 2026-03-02). Free does NOT include Claude Code, Research mode, full Opus access, or Microsoft 365 integration.
The real cost calculation for Free: Claude Sonnet 4.6 holds an AA-Omniscience hallucination rate of approximately 38% per Suprmind’s AI Hallucination Rates and Benchmarks reference – less than half GPT-5.2’s roughly 78%. For casual queries and exploratory work, Free is a meaningful tool. For any work where citation accuracy matters, enable web search before relying on responses.
Claude Pro: $20/month
Pro at $20/month ($17/month annual, $204/year) is the entry point for serious use. It runs Sonnet 4.6 by default with limited Opus 4.7 access available, more usage than Free, Research mode, unlimited Projects, Microsoft 365 integration (Excel, Outlook, Word – the last via the April 2026 launch), voice mode, and the option for extended thinking on Sonnet. Annual billing reduces the per-month rate approximately 17% to $17/month.
The Pro Claude Code question is genuinely conflicting as of 2026-05-07. The current anthropic.com/pricing page lists “Includes Claude Code” under Pro, and a Claude Code-specific pricing page describes “Claude Pro ($20/mo) for moderate use” as a path. An independent changelog tracker (scriptbyai.com, April 2026) states “Anthropic has removed Claude Code from its $20-a-month pro subscription.” These two sources disagree. The practical recommendation: confirm Claude Code inclusion on anthropic.com/pricing immediately before subscribing if Claude Code is the reason you’re considering Pro.
See also: Claude Code feature deep dive →
Claude Max 5x and Max 20x: $100/$200/month
Max 5x at $100/month and Max 20x at $200/month sit at the top of the consumer ladder. Both include the same model lineup as Pro plus early-access features and priority routing during high-traffic periods. Max 5x provides 5x more usage than Pro; Max 20x provides 20x more usage than Pro. Annual pricing for Max plans is not publicly disclosed in reviewed sources.
For developers and power users running Claude for hours per day on consequential tasks, Max 20x is the only tier where Opus usage does not consistently hit daily limits in practice. Per Anthropic’s documented rate-limiting policy, Max plan users report priority routing during overload periods – relevant after the 14-hour March 2-3 2026 outage that brought down claude.ai and the mobile apps while leaving the API largely functional.
Claude Team Standard and Team Premium
Team Standard at $25/seat/month ($20/seat/month billed annually) is for organizations with 5 to 150 seats. It bundles SAML SSO, central billing, admin controls, Microsoft 365 integration, no model training on team conversations by default, and the same model access as Pro. Team Premium at $125/seat/month ($100/seat/month billed annually) provides 5x the usage of Team Standard seats. Team is the right tier for organizations that need shared Projects and admin tooling but not full enterprise compliance infrastructure.
Note on user limit: Team plans require a minimum of 5 seats. Beyond 150 seats, Anthropic routes the conversation to Enterprise sales-assisted plans.
Claude Enterprise
Enterprise pricing starts at $20+/seat plus API usage and is annual only. Enterprise adds SCIM provisioning, audit logs, compliance API, custom data retention, HIPAA-ready offering (listed as “available” not “certified”), IP allowlisting, and a 500K token chat context window on some models (versus 200K on other plans). Enterprise has two paths: a self-serve tier with no published seat minimum, and a sales-assisted tier with custom terms, tiered usage incentives, and customer success support.
Enterprise data is contractually excluded from model training without per-user opt-out. This is the structural alternative to the Free, Pro, and Max plans where users must individually opt out of training (the August 2025 policy change made opt-out the default behavior; data retention extends to 5 years for users who do not opt out).
See also: Suprmind Multi-Model Divergence Index →
Anthropic API Pricing for Developers
API pricing is metered per million tokens with separate input, cached input write, cached input read, and output rates. Cached inputs (a request reusing prompt material from a recent prior request) get a 90% discount on read.
Source: anthropic.com/pricing, accessed 2026-05-07.
Additional API charges:
Batch API: 50% discount on all models. Opus drops to $2.50/$12.50, Sonnet to $1.50/$7.50, Haiku to $0.50/$2.50. Batch supports up to 10,000 queries for async processing in under 24 hours – the right path for high-volume non-interactive workloads where 24-hour latency is acceptable.
Per the Suprmind Multi-Model Divergence Index (April 2026, n=1,324 production turns), Claude’s catch ratio is 2.25 – it caught 304 errors made by other models and was caught 135 times. Combined with Perplexity (catch ratio 2.54), the two providers account for 60.7% of all corrections in the study. For verification-layer workflows where catching errors matters more than answering breadth, Claude API at $5/$25 (Opus) or $3/$15 (Sonnet) is competitive with peer pricing despite not being the cheapest option.
See also: AI catch ratio data →
Recent Pricing Changes (2025-2026)
Source: anthropic.com/pricing; metacto.com; pecollective.com.
The Opus 4.6 launch eliminated the long-context surcharge that had existed across the AI industry and dropped the Opus tier price 67%. Both effects compound: workloads that previously required either short context with Opus or long context at premium pricing now get full 1M context at the new lower base rate.
Geographic Availability and Restrictions
Both API and Claude.ai cover all EU member states, UK, US, Canada, Australia, Japan, India, South Korea, Brazil, most of Africa, Middle East, and Southeast Asia. Russia, China (mainland), North Korea, Iran, Cuba, Belarus, and the occupied Ukrainian territories of Crimea, Donetsk, Kherson, Luhansk, and Zaporizhzhia are not supported. Hong Kong is not on the supported list as of data reviewed.
EU data residency is available for the Anthropic API via multi-region processing since August 2025. Claude.ai consumer plans do not offer EU data residency by default – inference routes to Anthropic’s servers regardless of Azure region selected via Microsoft Azure AI Foundry. Microsoft Foundry EU support is listed as “Coming 2026” on Anthropic’s regional compliance page. Full EU AI Act high-risk system obligations apply from August 2, 2026, creating a compliance window that EU enterprise buyers should plan around.
Distribution platforms: Anthropic API direct, Amazon Bedrock, Google Cloud Vertex AI, Microsoft Azure AI Foundry (generally available for select models as of 2026, EU inference Coming 2026).
See also: Suprmind’s AI Hallucination Rates and Benchmarks reference →
What’s Included on Each Tier
Source: anthropic.com/pricing, May 2026.
Sources
- anthropic.com/pricing (canonical pricing source)
- anthropic.com/supported-countries
- platform.claude.com (API documentation)
- metacto.com (API pricing breakdown)
- pecollective.com (pricing history)
- Suprmind Multi-Model Divergence Index (catch ratio data)
- Suprmind AI Hallucination Rates and Benchmarks (per-model hallucination data)
- AWS Bedrock Anthropic model cards (legacy model EOL dates)
- Microsoft Learn (Foundry EU support timeline)
Last verified 2026-05-07.
FAQ
Frequently Asked Questions
Is Claude free in 2026?
Yes. The Free tier at $0 provides access to Claude Sonnet 4.6 (default) and limited Haiku 4.5 at usage limits described as “conversation budget” without specific counts. Free includes web search (toggleable), basic file uploads, Artifacts, and chat memory (since 2026-03-02). Claude Code, Research mode, and full Opus access require paid tiers.
How much does Claude Pro cost and is it worth it?
Pro is $20/month ($17/month annual) and is the entry tier most users should choose if they need more than casual access. It includes Sonnet 4.6 default with limited Opus 4.7 access, Research mode, unlimited Projects, Microsoft 365 integration, voice mode, and Cowork. Annual billing is $204/year, a roughly 17% reduction. Claude Code inclusion in Pro is contested – verify at anthropic.com/pricing before subscribing if Claude Code is the reason.
What is Claude Max and is it worth it?
Max comes in two variants: Max 5x at $100/month (5x Pro usage) and Max 20x at $200/month (20x Pro usage). Both include full Opus 4.7 access with extended thinking, priority routing during high-traffic periods, and early-access features. For developers and power users running Claude for hours per day on consequential tasks, Max 20x is the only tier where Opus usage does not consistently hit daily limits in practice.
What is the difference between Claude Team Standard and Team Premium?
Team Standard at $25/seat/month ($20/seat/month annual) requires a minimum of 5 seats and includes SAML SSO, central billing, admin controls, Microsoft 365 integration, and no model training on team conversations by default. Team Premium at $125/seat/month ($100/seat/month annual) provides 5x the usage of Team Standard seats with the same feature set.
What is the difference between Claude Team and Enterprise?
Team Standard is $25/seat/month (or $20/seat/month annual) for organizations with 5+ users. Enterprise adds SCIM provisioning, audit logs, compliance API, custom data retention, HIPAA-ready offering, IP allowlisting, and 500K context window on some models. Enterprise is appropriate for regulated-industry deployments; Team is for smaller organizations that need shared Projects and admin tooling but not full compliance infrastructure.
How much does the Claude API cost?
Claude Opus 4.7 API costs $5.00 per million input tokens, $0.50 per million cached input read, and $25.00 per million output tokens. Sonnet 4.6 costs $3.00/$15.00. Haiku 4.5 costs $1.00/$5.00. Batch API applies a 50% discount across all models. Claude Opus 4.1 (legacy, AWS Bedrock EOL 2026-05-31) was priced at $15/$75 – the 67% reduction at Opus 4.6 launch is the most significant pricing event in Claude’s API history.
Has Claude pricing changed recently?
Yes, significantly down. Claude Opus API pricing dropped from $15/$75 per million input/output tokens (Opus 4.1, mid-2025) to $5/$25 (Opus 4.6/4.7) – a 67% input price reduction and 67% output price reduction. The 1M context window also became standard at no surcharge. Sonnet 4.6 pricing remained unchanged at $3/$15. Haiku 4.5 introduced a new $1/$5 price point in October 2025.
What is Claude cached input pricing?
Cached input read costs 90% less than fresh input across the Claude 4 family ($0.50 vs $5.00 for Opus, $0.30 vs $3.00 for Sonnet, $0.10 vs $1.00 for Haiku). Cached input write costs 25% more than fresh input. The API automatically caches frequently used prompt prefixes when a request reuses substantial portions of recent prior prompts (default 5-minute TTL, extended TTL available). For workflows with stable system prompts or templated requests, caching is a 10x cost reduction with no engineering work.
Why was the Pro tier Claude Code change controversial?
The status of Claude Code in the $20 Pro tier is genuinely conflicting as of 2026-05-07. anthropic.com/pricing lists “Includes Claude Code” under Pro and a Claude Code-specific pricing page references Pro as a usage path. An independent changelog tracker (scriptbyai.com, April 2026) states Anthropic removed Claude Code from Pro. The conflict is unresolved in published sources. Verify directly at anthropic.com/pricing.
Are there annual discounts on Claude consumer tiers?
Pro is $17/month when billed annually ($204/year), an approximately 17% reduction from monthly. Team Standard is $20/seat/month annual versus $25 monthly. Team Premium is $100/seat/month annual versus $125 monthly. Annual pricing for Max 5x and Max 20x is not publicly disclosed in reviewed sources. Enterprise is annual-only.
How do I opt out of Anthropic using my conversations to train Claude?
Settings → Privacy → Data Usage → toggle off training consent. Anthropic changed from opt-in to opt-out by default in August 2025, extending data retention from 30 days to 5 years for users who do not opt out. Enterprise and Team plans include contract-level data non-training provisions without per-user opt-out.
Stop guessing. Start cross-checking.
Suprmind runs your prompt across ChatGPT, Claude, Gemini, Grok, and Perplexity in parallel. See where they agree, where they disagree, and which insights only one model surfaced — before you act.