Claude vs ChatGPT Enterprise for UK SMEs: a 2026 buyer's guide

What changed in the last 12 months

Both platforms have shipped enough material change in the last year that a 2024 or early-2025 comparison is out of date. OpenAI launched UK data residency for ChatGPT Enterprise, ChatGPT Edu, and the API Platform on 24 October 2025, in conjunction with the Stargate UK partnership. UK customers on eligible tiers can now store data at rest in the UK directly, rather than routing through Azure.

Anthropic made three significant moves. Claude Opus 4.6 and Sonnet 4.6 were released in February 2026, with the 1M-token context window moving to standard per-token pricing in March 2026 rather than carrying a premium surcharge above 200k tokens. The Claude Partner Network launched on 12 March 2026 with $100m in committed investment and Accenture, Deloitte, Cognizant, and Infosys as anchor partners. On 27 January 2026, Anthropic signed a partnership with DSIT to build AI assistance into GOV.UK services, signalling UK government preference for Claude as a delivery platform. Anthropic's own compliance certifications now include ISO/IEC 42001:2023 alongside SOC 2 Type II, ISO 27001:2022, HIPAA readiness, and UK Cyber Essentials.

Both platforms signed the EU General-Purpose AI Code of Practice in July 2025, which is relevant to UK SMEs serving EU customers. Both offer Data Processing Addenda on business and enterprise tiers.

Pricing in GBP, with VAT

Headline pricing for UK SMEs, exclusive of VAT unless noted, as of April 2026:

Claude. Claude Pro is around £16 per month for individuals. Claude Max runs around £80 per month. Claude Team is approximately £24 per seat per month on annual billing, with a five-user minimum. Claude Enterprise starts at around $30 per user per month with custom pricing above, and this is where the single most important commercial nuance sits, which we address below. VAT at 20% applies for UK business buyers.

ChatGPT. ChatGPT Plus is commonly quoted at around £20 per month inclusive of VAT at the UK checkout. ChatGPT Business (renamed from Team in August 2025) is £25 per user per month on annual billing, £30 monthly. ChatGPT Enterprise is custom-priced with a minimum 150-user commitment, typically estimated by industry analysts in the $40 to $60 per user per month range.

API pricing (token-based, April 2026). Claude Haiku 4.5 is $1 per million input tokens and $5 per million output. Claude Sonnet 4.6 is $3 in, $15 out. Claude Opus 4.6 is $15 in, $75 out. OpenAI GPT-5 nano is $0.05 in, $0.40 out. GPT-5 mini is $0.25 in, $2 out. GPT-5.3 is $1.75 in, $14 out. For high-volume automation workloads where the mini tier is adequate, OpenAI is materially cheaper per token. For tasks requiring long-context reasoning or frontier coding capability, Claude's flagship carries a premium but ranks higher on long-context benchmarks.

The Claude Enterprise pricing model deserves its own paragraph. Industry reports and community threads suggest the Enterprise seat fee includes no bundled token usage, with all API consumption billed on top at the rates above. This is different from Claude Team, where usage is included in the seat fee. The practical implication for UK SMEs evaluating Claude Enterprise: a detailed usage estimate is required before commit, because token costs at 50 or 100 users can materially exceed the seat fee. We recommend running a one-month Teams-tier pilot with usage logged, then using that usage profile to model Enterprise total cost. Anthropic's partner channel or direct sales team should confirm the current Enterprise structure before contract.

Total cost of ownership at 10, 25, and 50 seats

Using conservative usage assumptions for SME knowledge-work (mix of chat, summarisation, drafting, and light agentic workflows), annual cost sketches look roughly as follows, excluding VAT.

At 10 seats, Claude Team runs around £2,880 per year. ChatGPT Business runs around £3,000. Differences below £150 per year are within procurement noise. Choose on compliance fit and tool ecosystem rather than seat price.

At 25 seats, Claude Team is around £7,200, ChatGPT Business around £7,500. Claude Enterprise is not viable below the user threshold at which Anthropic will write a contract, typically 50+ seats. ChatGPT Enterprise requires 150 users minimum, so neither Enterprise tier is usually available.

At 50 seats, Claude Team is around £14,400 per year. Claude Enterprise at $30 per seat per month is around $18,000 in seat fees, roughly £14,300, plus API consumption layered on top. API consumption varies wildly: a team making heavy use of Claude Projects and long-context analysis can add £10,000 to £30,000 a year; a team running Claude for occasional drafting might add under £3,000. ChatGPT Enterprise requires a 150-user minimum, so it is not usually available at 50 seats.

The point of these sketches is not precision. It is that seat price alone is not a reliable comparison at Enterprise tier. Anyone signing Claude Enterprise without a usage-based cost model is exposed to bill surprise in month three.

UK data residency and GDPR posture

UK data residency was the single biggest compliance gap in early 2025 and is now partly closed. OpenAI's October 2025 launch means ChatGPT Enterprise, ChatGPT Edu, and eligible API Platform projects can store UK data at rest in the UK. This is a material advantage for regulated-sector SMEs, financial services, and any business whose clients require UK-only processing.

Claude data residency as of April 2026 is more nuanced. The Claude API supports EU and UK-eligible processing via AWS Bedrock in UK South or EU Ireland, and via Google Cloud Vertex AI in EU regions. The DEFRA AI SDLC tool guidance published in June 2025 confirms Bedrock UK regions as compliant for UK government data, which sets a useful precedent. However, the direct claude.ai web and desktop interface is still processed primarily in the United States at the time of writing. For UK SMEs in regulated sectors handling UK or EU personal data, the web interface is a compliance gap, and the operational answer is either the API via Bedrock or ChatGPT Enterprise on UK residency.

A Data Processing Addendum is required under UK GDPR for any AI vendor processing personal data, and both Anthropic and OpenAI offer DPAs on business and enterprise tiers. Consumer and free-tier accounts carry no DPA protection. A Data Protection Impact Assessment may be required for customer-facing or HR use cases regardless of which vendor is chosen.

Tooling and integrations

Claude's differentiator in 2026 is the depth of its agentic tooling. Claude Code is now bundled into Team and Enterprise seats, Computer Use has been integrated into Claude Code, and the Model Context Protocol gives Claude standardised connectors to a growing list of business applications, including Google Workspace as of December 2025 and the Microsoft 365 connector extended to all user plans in April 2026. Claude Projects and Artifacts support long-context document work well.

ChatGPT's differentiator is breadth. Custom GPTs, the Connectors framework, Canvas, and a larger third-party plugin ecosystem suit businesses that want to spin up many lightweight automations. DALL-E image generation and real-time voice ship in the product directly. For non-technical SMEs whose AI use case is messaging-style automation and content creation across many employees, ChatGPT's surface area is broader.

The practical view: if your business is coding-heavy, document-heavy, or needs deep agentic workflows, Claude is generally the stronger technical fit. If your business has many casual users needing a broad toolkit, ChatGPT Enterprise fits more naturally. Most enterprises run both, which matches the A16Z survey finding that 81% of Global 2000 companies use three or more AI model families concurrently.

UK-specific decision factors

Three factors matter specifically to UK buyers. First, UK data residency, already covered above. Second, Cyber Essentials. Both Anthropic and OpenAI hold UK Cyber Essentials certification, which matters for UK public-sector supply chains and some regulated buyers. Third, the Anthropic-DSIT partnership announced in January 2026 signals UK government preference for Claude in specific use cases, which is relevant to any UK SME bidding into public-sector work.

British English handling is a minor but real consideration. Independent testing has shown that LLMs trained predominantly on American English corpora tend to favour American spellings unless explicitly prompted otherwise. This is a configuration issue rather than a vendor-choice issue: both platforms respect UK English when the system prompt is set correctly, and both will drift without it.

VAT treatment is worth getting right at procurement. UK VAT at 20% applies to both subscriptions for business customers. VAT-registered businesses can reclaim input VAT in the normal way. Claude Pro is commonly quoted exclusive of VAT (£16 net, £19.20 gross). ChatGPT Plus is commonly quoted inclusive of VAT at the UK checkout (around £20 gross). Budget comparisons at the 10 to 50 seat level should normalise both to the same convention before being presented to finance.

What a sceptical CFO will push back on

Benchmark scores do not translate to business outcomes. Claude Opus 4.6 leads on long-context benchmarks such as MRCR v2 at the 1M-token length; GPT-5 leads on other reasoning tasks. Neither advantage is decisive for the majority of UK SME use cases, which are drafting, summarisation, customer service, and light automation. Pilot on your actual tasks, not on the published benchmarks.

The enterprise label is not a legal certification. Many UK SMEs buy Business or Team tiers and treat them as enterprise-grade without configuring admin controls, signing a DPA, or documenting usage policies. The plan tier is necessary but not sufficient for GDPR compliance. Whichever platform is chosen, the internal governance work still has to be done.

Vendor lock-in risk is real. Custom GPTs on OpenAI and Claude Projects and MCP integrations on Anthropic both create switching costs over time. Ask the commercial question now: what is the exit cost at 18 months if our preferred platform changes? A multi-vendor posture, or a deliberate choice to keep proprietary prompts and data portable, reduces this exposure.

Questions to answer before signing

Five questions sharpen the procurement decision:

1. What is our projected total cost at the end of year two, including token usage if Enterprise is on the table?
2. Does our sector or client base require UK data residency at rest, and if so which platform supports that natively today?
3. What is our dominant use-case profile, and which platform's tooling fits that profile best?
4. What admin controls, DPA, and usage policies will we put in place before rollout, regardless of platform?
5. What is our exit plan if we need to migrate vendors in 18 to 24 months?

The AI Consultancy runs procurement reviews for UK SMEs evaluating Claude, ChatGPT Enterprise, or a multi-vendor posture. For a readiness assessment before commit, see our AI readiness assessment. For the arithmetic on projected savings, see our AI ROI calculator.