Blog · Published · Updated
Local LLM or OpenAI, Claude and Gemini?
Short answer
Choose your own model when data must not leave the company or the load is large and steady. Choose the OpenAI, Claude or Gemini API for irregular traffic. In the scenario a GPU server with upkeep costs about €493 a month. The same amount buys about 30,700 documents a month on the Claude Sonnet 5.5 API, and twenty times more on Haiku 5.5.
How we workDetails ↓
- 01Call30 minutes with an engineer, free
- 02PrototypeA working system on your process and data, free
- 03ImplementationTarget and fixed price in the contract; 50% at signing
- 04Acceptance50% once the agreed target is met
- analysis with an engineer and a free prototype
- €0
- team in ChatGPT or Claude, up to 10 people
- from €1,150
- automation, agent or application for one process
- from €3,000
We belong to the OpenAI and Anthropic partner programmes and learn through their official courses. What this means ↓
Author: Syntalith · 3 min read
Statistics
AI models in companies: statistics 2026
30,700
documents a month on the Claude Sonnet 5.5 API cost as much as a GPU server with upkeep. Below that volume the API is cheaper. Scenario.
40%
According to Menlo Ventures (2025), Anthropic holds 40% of large companies' spending on language model APIs and OpenAI 27%, based on a survey of about 500 US decision-makers.
Source: Menlo Ventures, State of Generative AI in the Enterprise (2025) ↗
48.8%
According to Eurostat (2025), 48.8% of EU companies that considered AI and did not use it named concerns about data protection and privacy.
52.7%
According to Eurostat (2025), 52.7% of EU companies used paid cloud computing services.
20.0%
According to Eurostat (2025), 20.0% of EU companies with at least 10 employees used AI.
Figures checked: 22 October 2026
Compare the same work
A company that sends a dozen requests a day buys something different from a team that processes documents all night. In the first case it pays for access to a model. In the second, the server's throughput and how fully it is used matter most.
| Question | Your own model | A managed API |
|---|---|---|
| Who maintains the model | Your team or a contractor | The service vendor |
| Where the input goes | To the environment you design | To the regions described in the service terms |
| How you change the model | Weights, configuration, server tests | Choosing an available model, app compatibility |
| How you pay | Hardware or rental, upkeep, energy | API usage, integration and quality control |
The line does not run between companies: OpenAI also publishes the open gpt-oss, and Google releases Gemma.
What each path costs
- Hetzner GEX45 GPU server
- €214 a month net
- Server upkeep
- 8 h (€279)
- Your own model in total
- €493 a month
- One document on the Claude Sonnet 5.5 API
- $0.018 (€0.016)
- One document on the Claude Haiku 5.5 API
- $0.0009 (€0.0008)
Break-even of the server against Sonnet 5.5
30,700 documents a month
Price of the GEX45 server with 24 GB of GPU memory from Hetzner's page on 22 October 2026, without the one-off €209 fee. €34.90 an hour is the EU average hourly labour cost for 2025 according to Eurostat. The API at Anthropic's rates, at the ECB rate of 1.1205 USD per euro on 22 October 2026. The upkeep time and the document size are the scenario's assumptions. The local model's quality must be comparable, which you check with a test.
A document: 5,000 input tokens and 800 output tokens. For comparison, a GPU server with upkeep costs €493.
Źródło: Anthropic, API pricing (2026) ↗
On Claude Haiku 5.5 the same document costs twenty times less, so the server's break-even rises to hundreds of thousands of documents. A local model rarely wins on price. It wins on the requirement that data stays inside the company.
Owning the hardware changes the count: you pay once for the server, then for energy, space and upkeep. The break-even then falls, but you take on the risk of failure and replacement, which the vendor carries when you rent.
Privacy: check the product and the settings
OpenAI states that API data does not train models by default, unless the customer opts in. Retention is a separate question. Anthropic describes no training by default in its commercial products. The Gemini API terms distinguish paid and free services and include regional rules. Your own server also needs rules for access and retention.
Where a local model wins
A production plant without internet access
a local model
priced after a call
A document policy forbids sending anything outside the company
a local model
priced after a call
A steady, large load around the clock
a cost comparison at your volume
priced after a call
Irregular traffic, a few hundred cases a day
API
by usage
Quality in Polish, or in any language other than English, needs its own trial, as in our Polish language test for local LLMs: negation, inflected names, missing information. You can also split the work: a local model classifies documents, and hard cases go to a person or an API, with an explicit rule on which data may leave. We prepare the comparison on your tasks as part of private LLMs.
Worked example · Scenario
How we count the payback: a local model in manufacturing
We run the same count for your volume in the free process review.
€5,817
a month for the work on 1,000 questions
EU cloud version payback: 0.6 mo.
Our assumptions are 10 minutes to find the right instruction, and checking in 15% of the old time. €34.90 an hour is the EU average hourly labour cost for 2025 according to Eurostat, for companies with 10 or more employees. Model cost: Claude Sonnet 5.5, about 6,000 + 500 tokens per answer in the EU cloud version, at the ECB rate of 1.1205 USD per euro on 22 October 2026. We count the payback for an app with a model in the EU cloud (from €3,000 net). We price the model on your server after a call, and the cost of people's work stays the same.
See the full calculation →
