Enterprise AI Platform Pricing - May 2026

Amazon Q Lite stays cheapest at $3/user, ChatGPT Business dropped to $20 annual, and Anthropic shifted Enterprise to pure usage-based billing.

Cheapest: Amazon Q Business Lite Best Value: ChatGPT Business Updated weekly
Enterprise AI Platform Pricing - May 2026

TL;DR

  • Amazon Q Business Lite at $3/user/month stays the cheapest enterprise AI seat
  • ChatGPT Business dropped to $20/user/month (annual) on April 2 - best mid-market value
  • Anthropic flipped Claude Enterprise to pure usage-based billing: $20/seat plus API tokens on top
  • Microsoft launched the M365 E7 bundle at $99/user/month on May 1, folding Copilot in

What Changed Since March 2026

A lot moved in eight weeks. OpenAI cut ChatGPT Business prices, Anthropic restructured Enterprise billing in a way that'll raise costs for heavy users, and Microsoft shipped a new top-tier bundle that finally makes Copilot a bundled default rather than an add-on.

The most consequential shift is Anthropic's. Enterprise customers renewing after November 2025, and all new Enterprise contracts from February 2026, now pay a flat $20/seat/month plus standard API rates for every token consumed. The old model bundled an usage allowance into the seat fee and offered a 10-15% discount off API rates. The new model does neither. For teams where API usage already dominated costs, the headline seat price drop from ~$60 to $20 is a mirage. The Register reported that for some customers the seat fee was only 20% of total spend, meaning the other 80% just got more expensive.

OpenAI moved in the opposite direction. On April 2, 2026, ChatGPT Business dropped $5 to $20/user/month (annual) or $25/month (monthly). OpenAI also introduced Codex-only seats - no fixed seat fee, just pay-as-you-go token consumption for teams that only need the coding agent.

Claude Team plans got a double boost in May: Standard seats dropped to $20/user/month, and rate limits across all paid Team plans doubled following Anthropic's compute expansion deal with SpaceX's Colossus 1 cluster.

Full Pricing Table

All prices in USD per user per month. Verified against official pricing pages on May 18, 2026.

PlatformPlanPrice/User/MonthMin SeatsBillingKey Inclusions
Amazon Q BusinessLite$31MonthlyBasic Q&A, permission-aware responses
Amazon Q BusinessPro$201MonthlyFull suite, Q Apps, QuickSight, Slack/Teams
ChatGPTBusiness$20 (annual) / $25 (monthly)2Annual or monthlyGPT-5.5, admin console, SSO, Codex access
ClaudeTeam Standard$20 (annual) / $25 (monthly)5Annual or monthlyClaude Opus 4.7, projects, SSO
ClaudeTeam Premium$100 (annual) / $125 (monthly)5Annual or monthlyEverything + Claude Code, 6.25x usage
Google WorkspaceBusiness Standard$141AnnualGemini in Gmail/Docs/Meet/Sheets, 2 TB
Google WorkspaceBusiness Plus$221AnnualGemini expanded, NotebookLM Plus, 5 TB
Microsoft 365 CopilotAdd-on$301AnnualCopilot in Word/Excel/PPT/Outlook/Teams
Microsoft 365E7 Bundle$991AnnualM365 E5 + Copilot + Entra Suite + Agent 365
PerplexityEnterprise Pro$40 (or $33.33 annual)1Annual or monthlyAI search, SSO, 500 research queries/month
PerplexityEnterprise Max$3251MonthlyAll models, 4000 Pro Searches/week, video gen
ClaudeEnterprise$20 + usageCustomAnnualFull Claude suite, HIPAA-ready (sales), SCIM
ChatGPTEnterprise~$60-100 (negotiated)150AnnualUnlimited GPT-5.5, HIPAA, SCIM, audit logs
Google WorkspaceEnterpriseCustom1AnnualEnterprise security, unlimited storage

For API-level pricing on the underlying models, see our LLM API pricing comparison.

ChatGPT Enterprise and Claude Enterprise don't publish fixed prices. The $60-100 figure for ChatGPT Enterprise comes from third-party analyses and user reports; actual rates vary by seat count and contract length. Claude Enterprise's $20/seat is the published rate, but the real cost depends entirely on how much your team uses the models. Both require annual contracts and sales conversations.

Two people reviewing enterprise software options on a laptop Picking an enterprise AI platform now requires reading the billing model as carefully as the feature list. Source: unsplash.com

Security and Compliance Comparison

Enterprise contracts are largely sold on compliance posture. The gap between a $20/month team plan and a $60-100/month enterprise contract often comes down to HIPAA eligibility and a Business Associate Agreement.

PlatformSOC 2 Type IIHIPAA/BAAGDPRSSOSCIMAudit LogsData Training Opt-Out
ChatGPT EnterpriseYesYes (BAA available)YesSAMLYesYesYes (default)
ChatGPT BusinessYesNoYesSAMLNoLimitedYes (default)
Claude EnterpriseYesYes (sales-assisted)YesSAML/OIDCYesYesYes (default)
Claude TeamYesNoYesSSONoLimitedYes (default)
Microsoft M365 CopilotYesYes (via M365)YesAzure AD/EntraYesYesYes
Google Workspace EnterpriseYesYes (via Workspace)YesGoogle SSOYesYesConfigurable
Amazon Q BusinessYes (via AWS)YesYesIAM/SAMLYesCloudTrailYes
Perplexity Enterprise ProYesNoYesSSO/MFANoAdmin panelYes
Perplexity Enterprise MaxYesNoYesSSO/MFAYes (50+ seats)Yes (webhook)Yes

HIPAA Availability in Practice

OpenAI offers BAA support at the ChatGPT Enterprise tier. HIPAA isn't available on Business or Team plans. Anthropic follows the same split: BAA is available for Enterprise (sales-assisted) only, not for Team Standard or Team Premium.

Microsoft and Google inherit their compliance posture from their cloud platforms. Organizations with an existing Microsoft 365 E3/E5 or Google Workspace Enterprise BAA extend that protection to Copilot or Gemini without a separate negotiation. That matters a lot if you're already in one of those ecosystems.

Amazon Q Business became HIPAA-eligible through AWS's compliance framework in October 2024. Existing AWS BAA holders have coverage for Q Business without additional paperwork.

Perplexity Enterprise Pro and Max are SOC 2 Type II certified and GDPR compliant, but neither tier offers HIPAA support. That rules them out for any healthcare or life sciences use case handling PHI.

Hidden Costs Most Comparisons Miss

Claude Enterprise's Usage Flip

This is the one most teams will get wrong in 2026. The new Claude Enterprise structure lists $20/seat/month, which sounds cheap. But every token consumed in Claude chat, Claude Code, and Cowork now bills at standard API rates with no discount. The old enterprise contract bundled usage and offered a 10-15% discount off standard rates. The new one does neither. For an engineering team running Claude Code heavily, Anthropic's own data shows developers average $150-250/month in consumption - meaning the real per-seat cost is closer to $170-270/month for a developer, not $20. That is 2-3x the old pricing for heavy users.

Signing a contract without reading the billing terms is the fastest way to overspend on enterprise AI The billing model matters as much as the per-seat price. Read the consumption terms before signing. Source: unsplash.com

Microsoft's Double Bill

The standalone Microsoft 365 Copilot add-on at $30/user/month sits on top of an existing qualifying M365 license. That's $30 + $22 (Business Standard) = $52/user, or $30 + $57 (E3) = $87/user before you even start using AI. The new M365 E7 bundle at $99/user/month bundles Copilot, Entra Suite, and Agent 365 on top of M365 E5, which is better value than buying those pieces separately at ~$117/user/month. But if you aren't already a M365 shop, none of this is cheap.

Amazon Q's Index Overhead

Amazon Q Business charges $3 or $20 per user, but the index is a separate line. A Starter index unit runs $0.14/hour (roughly $100/month), and an Enterprise index unit costs $0.264/hour (roughly $190/month). Media processing adds $0.003 per image and $0.05 per minute of video. For 100 seats with three Enterprise index units, index costs alone add about $570/month to the bill.

ChatGPT Codex Seats Have No Cap

The new Codex-only seat has no fixed monthly fee, which sounds economical. But there's also no built-in rate limit, and token consumption for automated coding tasks can spike fast. OpenAI recommends setting workspace credit budgets if you're adding Codex seats at scale. Without a budget cap, a few engineers running large repo scans can push the monthly bill well past what a fixed-price seat would have cost.

Perplexity Enterprise Max at $325/user/month is a new tier that most of the comparisons haven't caught up with yet. It bundles 4,000 Pro Searches per week, 500 Deep Research tasks per month, video generation, and SCIM provisioning.

Google's Bundled Price Increase

Google stopped selling standalone "Gemini Enterprise" add-ons for new customers in early 2026. Gemini is now baked into Business Standard ($14/user/month), Business Plus ($22/user/month), and Enterprise (custom) plan prices, which all went up by roughly $2-4/user/month to absorb the AI cost. Existing customers on legacy Gemini add-ons keep their rates until renewal.

Decision Guide

Choose ChatGPT Business ($20/user/month, annual) if:

You want broad model access (GPT-5.5, o3, DALL-E, Codex) with admin controls at the tightest mid-market price. Works for general-purpose teams. No HIPAA, but SOC 2 covered. The two-seat minimum makes it accessible even for small teams, and the April 2026 price drop makes it the clearest value at this tier. See our ChatGPT review for how it holds up in day-to-day use.

Choose Claude Team Premium ($100/user/month, annual) if:

Your team needs heavy coding throughput and doesn't mind the usage caps. The May 2026 rate limit doubling made the Premium seat meaningfully better than it was six months ago. Claude Code access is the draw. For how Claude performs on real coding tasks, see our Claude Sonnet 4.6 review.

Choose Claude Enterprise ($20/seat + usage) if:

You need Anthropic's full compliance stack - HIPAA/BAA via sales, SAML/OIDC SSO, SCIM, audit logs. Budget for the usage costs honestly. If your team is research-heavy rather than coding-heavy, consumption will be lower than the Claude Code estimates above. For teams building AI agents, Enterprise unlocks the full API integration layer.

Choose Microsoft 365 E7 ($99/user/month) if:

Your organization is already paying for Microsoft 365 E5 and considering Copilot. The E7 bundle replaces ~$117 in separate licenses at $99, and it includes Agent 365 and Entra Suite that most enterprise IT teams will want anyway. If you're not already on E5, this isn't the entry point.

Choose Amazon Q Business ($3-20/user/month) if:

You're running on AWS and want AI that's aware of your internal permissions and data. The Lite tier at $3/user does genuine work for knowledge retrieval. Pro at $20 adds plugin integrations and Q Apps. HIPAA-eligible through AWS's framework. Watch the index costs before committing.

Choose Perplexity Enterprise Pro ($40/user/month) if:

Your primary use case is research and cited answers, not general-purpose AI. Perplexity excels at web-grounded answers with sourcing. SOC 2 certified, no HIPAA. The new Enterprise Max at $325/user is for teams that need video generation and very high research volume alongside the standard search capabilities. See our Perplexity review for real-world context.

Price History

  • May 2026 - Claude Team Standard drops to $20/seat/month (annual). Rate limits double for all paid Team plans following Anthropic's compute expansion. Microsoft 365 E7 bundle launched at $99/user/month.

  • Apr 2026 - ChatGPT Business drops $5 to $20/seat/month (annual). OpenAI adds Codex-only pay-as-you-go seats for Business and Enterprise workspaces. GPT-5.5 becomes top model on Business and Enterprise.

  • Feb 2026 - Claude Enterprise switches to $20/seat + usage billing by default for all new contracts. Token bundle discounts removed.

  • Jan 2026 - Anthropic cuts Claude Team Premium from $150 to $100/seat/month. OpenAI launches ChatGPT for Healthcare with separate BAA support. Google bundles Gemini into Workspace plans with $2-4/user price increases.

  • Nov 2025 - Anthropic begins renewing Enterprise customers under usage-based plans instead of flat-rate bundles. Microsoft announces M365 E7 bundle (available May 2026).

  • Oct 2024 - Amazon Q Business becomes HIPAA-eligible through AWS compliance framework.

Enterprise teams evaluating AI platforms must factor in both compliance requirements and total cost of ownership Enterprise AI buying decisions in 2026 increasingly involve procurement, legal, and finance - not just the engineering team. Source: unsplash.com

FAQ

Which enterprise AI platform is cheapest per seat?

Amazon Q Business Lite at $3/user/month is the lowest. For a full-featured platform, ChatGPT Business and Claude Team Standard both start at $20/user/month on annual billing as of May 2026.

Is Claude Enterprise actually $20 per user now?

The seat fee is $20/user/month, but every token consumed in chat, Claude Code, and Cowork bills on top at standard API rates. For developers using Claude Code heavily, total cost runs $170-270/user/month based on Anthropic's own usage data.

Which platforms support HIPAA with a BAA?

ChatGPT Enterprise, Claude Enterprise (sales-assisted only), Microsoft 365 Copilot (via your M365 BAA), and Amazon Q Business. Perplexity Enterprise Pro and Max, Claude Team, and ChatGPT Business don't offer BAA support.

Does Microsoft Copilot pricing include Microsoft 365?

No. The $30/user/month Copilot license is an add-on to a qualifying Microsoft 365 subscription. The new M365 E7 bundle at $99/user/month does include Copilot, Entra Suite, and Agent 365 on top of M365 E5.

What is Perplexity Enterprise Max?

A $325/user/month tier launched in 2026. It adds 4,000 Pro Searches per week, 500 Deep Research tasks per month, video generation (15 videos monthly), SCIM provisioning, webhook-delivered audit logs, and configurable data retention. No seat minimum.

How do ChatGPT Business and Claude Team compare at the same price?

Both now sit at $20/user/month (annual). ChatGPT Business includes unlimited GPT-5.5 use within plan limits with no per-token billing. Claude Team Standard includes the Claude app with an usage cap, but no Claude Code unless you add Premium seats at $100/user. For coding-heavy teams, the structures are very different.


Sources:

✓ Last verified May 18, 2026

LinkedIn
Reddit
Hacker News
Telegram
Enterprise AI Platform Pricing - May 2026
About the author AI Benchmarks & Tools Analyst

James is a software engineer turned tech writer who spent six years building backend systems at a fintech startup in Chicago before pivoting to full-time analysis of AI tools and infrastructure.