Updated March 2026
Claude Enterprise Pricing
Volume discounts, SSO, audit logging, custom rate limits, and dedicated support. Enterprise plans are negotiated directly with Anthropic based on your requirements and expected usage.
Contact Anthropic SalesWhat Claude Enterprise Includes
Claude Enterprise is designed for organisations that need security controls, compliance features, and guaranteed performance beyond what the standard API provides. All standard API features (prompt caching, Batch API, streaming) remain available, with additional enterprise-grade capabilities.
Single Sign-On (SSO)
SAML 2.0 and OIDC integration with your identity provider. Enforce access policies and automate user provisioning through your existing IdP.
Audit logs
Comprehensive audit trail for all API activity. Track who accessed what, when, and from where. Export logs to your SIEM for compliance and security monitoring.
Higher rate limits
Significantly increased requests per minute, tokens per minute, and tokens per day limits beyond the standard Scale tier. Custom limits negotiated based on your expected workload.
Dedicated support
Named account manager, priority support queue, and SLA guarantees. Direct engineering escalation for production issues.
Data retention controls
Zero data retention option where your prompts and responses are not stored by Anthropic beyond the lifetime of the request. Configure retention policies to meet regulatory requirements.
Custom model access
Early access to new models and capabilities. Option to work with Anthropic on custom model configurations or evaluations.
Standard API Pricing (Baseline)
Enterprise pricing starts from these standard rates. Volume discounts reduce these costs based on monthly commitment.
| Model | Input / MTok | Output / MTok | Context |
|---|---|---|---|
| Claude Opus 4 | $15.00 | $75.00 | 200K |
| Claude Sonnet 4 | $3.00 | $15.00 | 200K |
| Claude Haiku 3.5 | $0.80 | $4.00 | 200K |
All standard cost optimisation features apply: prompt caching (90% off input), Batch API (50% off all tokens), and streaming. Enterprise volume discounts are applied on top of these savings, not instead of them.
Volume Discounts
How volume pricing works
Anthropic does not publish a fixed volume discount schedule. Discounts are negotiated based on several factors:
- -Monthly committed spend (typically $10K+ for meaningful discounts)
- -Contract length (annual commitments get better rates)
- -Model mix (heavy Opus usage vs Haiku-dominant workloads)
- -Pre-payment willingness (upfront payment often unlocks deeper discounts)
- -Strategic partnership potential
- -Competitive situation (switching from another provider)
Annual commitment benefits
Enterprise contracts with annual commitments typically include:
Guaranteed rate locks
Your per-token rates are locked for the contract duration, protecting against price increases.
Volume discount tiers
The more you commit to, the deeper the discount. Typical range is 10-30% off standard pricing.
Reserved capacity
Guaranteed throughput allocation so your rate limits are not affected by platform-wide demand spikes.
Overage flexibility
Usage above your committed amount is billed at standard rates, not at penalty rates. No cliff fees.
Claude via Cloud Platforms
Claude is available through AWS Bedrock and Google Cloud Vertex AI. Pricing is broadly comparable to the direct API, but each platform has its own billing model.
| Provider | Sonnet Input/MTok | Sonnet Output/MTok | Billing |
|---|---|---|---|
| Direct Anthropic API | $3.00 | $15.00 | Pay-as-you-go or committed use |
| AWS Bedrock | ~$3.00 | ~$15.00 | AWS billing, reserved capacity available |
| Google Cloud Vertex AI | ~$3.00 | ~$15.00 | GCP billing, committed use discounts |
When AWS Bedrock makes sense
- -You already run infrastructure on AWS and want consolidated billing
- -Your data must stay within your AWS VPC for compliance reasons
- -You want to use AWS IAM for access control instead of Anthropic API keys
- -You have existing AWS reserved instance or savings plan credits
When Google Vertex AI makes sense
- -You are a GCP-first organisation with existing committed use discounts
- -You want to combine Claude with other Vertex AI services (Gemini, custom models)
- -Your compliance requirements mandate GCP infrastructure
- -You need multi-model orchestration within a single platform
Claude Enterprise vs OpenAI Enterprise
A feature-by-feature comparison for organisations evaluating enterprise AI platforms.
| Feature | Claude (Anthropic) | OpenAI |
|---|---|---|
| Context window | 200K tokens (all models) | 128K tokens (GPT-4o), 200K (o1) |
| SSO / SCIM | Yes (Enterprise plan) | Yes (Enterprise plan) |
| Data retention | Zero retention option | Zero retention option |
| Audit logs | Yes | Yes |
| Prompt caching | Yes (90% input savings) | Yes (50% input savings) |
| Batch processing | 50% discount, 24h window | 50% discount, 24h window |
| Rate limits | Custom (Scale/Enterprise) | Custom (Tier 5+) |
| Agentic coding | Claude Code (Opus/Sonnet) | Codex (o3/o4-mini) |
| Mid-tier pricing (input) | $3.00/MTok (Sonnet 4) | $2.50/MTok (GPT-4o) |
| Budget tier pricing (input) | $0.80/MTok (Haiku) | $0.15/MTok (GPT-4o-mini) |
The practical difference
Both platforms offer comparable enterprise features. The decision typically comes down to model quality for your specific use case, existing cloud relationships (Bedrock vs Azure OpenAI), and pricing negotiations. Many enterprise customers use both providers, routing different tasks to whichever model performs best. Claude typically leads on coding quality, long-context tasks, and instruction following, while OpenAI has advantages in reasoning (o1 family) and offers a cheaper budget tier (GPT-4o-mini).
How to Get Enterprise Pricing
Estimate your usage
Before contacting sales, estimate your monthly token volume, model mix (which models you will use most), and whether your workload fits the Batch API or prompt caching. This helps Anthropic provide an accurate quote.
Contact Anthropic sales
Reach out via anthropic.com/contact-sales or email [email protected]. Include your company name, expected monthly spend, primary use cases, and any compliance requirements (HIPAA, SOC 2, GDPR).
Technical evaluation
Anthropic will typically provide a trial period with increased rate limits so you can evaluate model quality on your actual workload. Use this time to benchmark against competitors.
Negotiate terms
Discuss contract length, committed spend tiers, rate limits, support SLAs, and data retention policies. Annual commitments with higher spend typically unlock the best per-token rates.
Deploy and optimise
Once your enterprise agreement is in place, implement prompt caching and Batch API to stack additional savings on top of your negotiated volume discount.