Mistral Large 4’s list price is $1.36 per million input tokens and $4.18 per million output tokens, with a cache price of $0.14. Sonnet 5.5 is $2 and $10, with a $0.10 cache read. Large 4 is the cheaper list price on a typical mix of input and output. A launch discount of 50 percent ran for two weeks from 6 October 2026 and had ended by the day these prices were checked, so the list price is the one used here. Large 4 is still marked as a public preview.
For 1,000 input tokens, 400 output tokens, and 50 requests a day, in US dollars: Mistral Large 4 is the cheapest of these, at $4.55 a month.
$0.003 for one request, and $0.15 a day at 50 requests a day.
Show the maths
New input: 1,000 tokens × $1.36 per million = $0.0014
Output: 400 tokens × $4.18 per million = $0.0017
1,500 requests a month × $0.003 = $4.55
Caching half the input would save about $0.92 a month
That uses the published cache-read price of $0.14 per million tokens. It assumes half of these input tokens are already in a cache. Writing the cache can cost extra, and that write price is not included.
Batch processing would save about $2.27 a month
Mistral says batch processing is half price. Batch jobs are slower, and they suit work that can wait.
Ministral 3 3B would cost about $4.34 less a month
It is a cheaper model from the same company. It is a different model, and the replies will not match.
A reply half as long would save about $1.25 a month
Output tokens are often the expensive part. Asking for a shorter answer cuts that cost directly.
$0.006 for one request, and $0.30 a day at 50 requests a day.
Show the maths
New input: 1,000 tokens × $2.00 per million = $0.002
Output: 400 tokens × $10.00 per million = $0.004
1,500 requests a month × $0.006 = $9.00
Caching half the input would save about $1.43 a month
That uses the published cache-read price of $0.10 per million tokens. It assumes half of these input tokens are already in a cache. Writing the cache can cost extra, and that write price is not included.
Batch processing would save about $4.50 a month
Anthropic's Batch API is half price on input and output. Cache discounts stack with that. Batch jobs are slower, and they suit work that can wait.
Claude Haiku 5.5 would cost about $8.55 less a month
It is a cheaper model from the same company. It is a different model, and the replies will not match.
A reply half as long would save about $3.00 a month
Output tokens are often the expensive part. Asking for a shorter answer cuts that cost directly.
Cached prices here are read prices. Writing a cache can cost extra, and that write price is on the provider’s page, not in this total. A US-only endpoint is 1.1× at some providers and is not included. Fast modes are not included.