Codestral vs GPT-5 mini

Codestral is cheaper for every workload modelled here.

List prices checked

Specifications and rates

PropertyCodestralGPT-5 mini
Input per 1M$0.300$0.250
Output per 1M$0.900$2.00
Cached input per 1M$0.025
Blended (3:1)$0.450$0.688
Context window262K400K
Max output128K
Inputs acceptedtexttext, image

Source: Mistral AI pricing (May 2026) · OpenAI pricing (May 2026). Published list prices only — committed-use discounts, regional variation, and per-request charges are not included. Not yet independently re-checked against the vendor’s page.

Monthly cost by workload

Identical token volumes through both rate cards. Output-heavy profiles are where the ranking flips.

WorkloadCodestralGPT-5 miniCheaper
In-app chat assistantA support or help assistant handling a few thousand conversations a month, with short prompts and short answers.$11$15Codestralsaves $4.50/mo (30%)
RAG over documentsRetrieval-augmented answering where each request stuffs several retrieved passages into the prompt, so input dominates.$104$105Codestralsaves $1.50/mo (1%)
Coding agentAn agentic loop that reads files and writes patches, generating heavy output and re-sending large context on every turn.$210$300Codestralsaves $90/mo (30%)

What each one is for

Codestral

by Mistral AI

A code-specialised model tuned for fill-in-the-middle completion, priced for the high request volume that inline editor autocomplete generates.

GPT-5 mini

by OpenAI

A fifth of GPT-5's price with the same 400k context window, which makes it the cheapest way to process very long documents without dropping to a small model.

Common questions

Is Codestral or GPT-5 mini cheaper in 2026?
Codestral, for every workload modelled here. On a blended 3:1 basis it costs $0.450 per million tokens against GPT-5 mini's $0.688.
Which has the larger context window?
GPT-5 mini, at 400K tokens against 262K. A larger window only helps if you can afford to fill it — the cost of one full-context request is in the table above.
Which is cheaper for output-heavy work like agents?
Codestral, at $0.900 per million output tokens against $2.00. Reasoning tokens bill as output, so for agentic loops this axis usually decides the invoice regardless of what the input rates look like.
What is not included in these figures?
Batch-processing discounts, fine-tuning, long-context surcharges above a provider's standard tier, and any enterprise agreement. These are published pay-as-you-go list prices, checked against each provider's own pricing page on the date shown.
Is switching from GPT-5 mini to Codestral worth it?
On price alone it is about 35% cheaper on a blended basis. Whether that survives contact with your workload depends on output quality and on retry rate — a cheaper model that needs two attempts is not cheaper. Model the specific volumes in the calculator before you migrate.

Other model comparisons