Cheapest EU and open-weight model APIs in 2026?

Mistral Small from Mistral AI — $0.15/1M blended. DeepSeek-V4-Flash is second at $0.18/1M blended. This ranking is specific to the workload described above; a different usage shape reorders it.

Why does this differ from the general cheapest list?

Because a single price axis never describes a real workload. Ranking by headline rate answers "who is cheapest per unit"; this page answers "who is cheapest for this job", and the two orders are often very different — which is the whole reason it exists as a separate table.

What is not accounted for?

Committed-use discounts, regional price variation, per-request charges, support plans, and minimum retention. These are published list prices applied to one stated workload — a shortlist to price properly with the vendor, not a quotation.

Cheapest EU and open-weight model APIs

Avoiding US hyperscaler lock-in, for data residency or licensing reasons.

List prices checked August 2026

How this is scored

Models from providers outside the big three US labs, ranked on blended token price. Some are open-weight and can be self-hosted later, which changes the negotiating position even if you start on the hosted API.

All 7, ranked for this workload

#	Model	Provider	Cost for this workload	Notes
1	Mistral Small	Mistral AI	$0.15/1M blended	An open-weight model available both as a hosted API and as a download, which makes it the cheap tier you can also run yourself if the economics change.
2	DeepSeek-V4-Flash	DeepSeek	$0.18/1M blended	A million-token context window at a fraction of any Western model's price, with cache-hit input billed at a fiftieth of the base rate. DeepSeek has published notice of a significant price rise, so treat these rates as a floor rather than a plan.
3	Cohere Command R	Cohere	$0.26/1M blended	The cheap end of Cohere's range, aimed at retrieval pipelines where the model's job is to ground an answer in supplied documents rather than to reason from scratch.
4	Codestral	Mistral AI	$0.45/1M blended	A code-specialised model tuned for fill-in-the-middle completion, priced for the high request volume that inline editor autocomplete generates.
5	DeepSeek-V4-Pro	DeepSeek	$0.54/1M blended	Frontier-class reasoning at roughly a fifth of Western frontier pricing, with a million-token window. The trade-offs are Chinese data residency, a 500-request concurrency cap, and a price rise the vendor has already announced.
6	Mistral Large	Mistral AI	$3.00/1M blended	The French lab's flagship, priced below every US frontier model and hosted in the EU, which is often the deciding factor rather than the benchmark scores.
7	Cohere Command A	Cohere	$4.38/1M blended	Cohere's enterprise flagship, built around retrieval-augmented generation and multilingual work, and sold heavily as a private deployment inside customer infrastructure.

Published list prices only. Rows that do not publish a rate for an axis this workload depends on are excluded rather than shown as cheap.

Questions

Cheapest EU and open-weight model APIs in 2026?: Mistral Small from Mistral AI — $0.15/1M blended. DeepSeek-V4-Flash is second at $0.18/1M blended. This ranking is specific to the workload described above; a different usage shape reorders it.
How is this ranking weighted?: Models from providers outside the big three US labs, ranked on blended token price. Some are open-weight and can be self-hosted later, which changes the negotiating position even if you start on the hosted API.
Why does this differ from the general cheapest list?: Because a single price axis never describes a real workload. Ranking by headline rate answers "who is cheapest per unit"; this page answers "who is cheapest for this job", and the two orders are often very different — which is the whole reason it exists as a separate table.
What is not accounted for?: Committed-use discounts, regional price variation, per-request charges, support plans, and minimum retention. These are published list prices applied to one stated workload — a shortlist to price properly with the vendor, not a quotation.

All 7, ranked for this workload

Questions

Other workloads