Text model

Devstral 2512 API

Devstral 2512 is a text model from Devstral, live on Corent. Answers in the OpenAI-compatible chat format.

Two ways to run it

Name Devstral 2512, or let Corent choose

You pick

POST /v1/chat/completions{  "model": "corent-devstral-2512",  "messages": [{ "role": "user",    "content": "Write a haiku about shipping fast." }]}

Direct access runs Devstral 2512 and no other model. No substitution, you pay its price plus Corent’s fee, and the receipt names the model you pinned.

The router picks

POST /v1/chat/completions{  "model": "corent",  "messages": [{ "role": "user",    "content": "Write a haiku about shipping fast." }]}

Leave the model out and Corent reads the prompt, weighs every live text model on quality and price, and picks Devstral 2512 whenever it is the best fit. The receipt shows the exact charge.

First call

One key, one request

curl -X POST https://api.corent.tech/v1/chat/completions \  -H "Authorization: Bearer $CORENT_API_KEY" \  -H "Content-Type: application/json" \  -d '{"model": "corent-devstral-2512", "messages": [{"role": "user", "content": "Write a haiku about shipping fast."}]}'

The response carries the result and a receipt: the model that served you and the exact cost in cents. Full request and response shapes are in the docs.

Details

Devstral 2512, as the API serves it today

Pin by namecorent-devstral-2512
Kind of resultText
EndpointPOST /v1/chat/completions
Made byDevstral
Quality4.4 out of 10, Corent’s score
StatusLive
Typical price$0.03 to $0.10 a typical reply. See pricing

These fields come from the live catalog, the same list the full model catalog prints, so they match what the API serves.

FAQ

Common questions

Does Devstral 2512 have an API?

Yes. Devstral 2512 is served through the Corent API on POST /v1/chat/completions. Pin the exact name corent-devstral-2512 and you get that model or nothing: no substitution, and the receipt names the model you pinned. One Corent key covers it alongside every other model in the catalog.

How much does Devstral 2512 cost on Corent?

You pay the model’s price plus Corent’s fee, and the exact charge comes back on every response. Text is billed per token, and a typical reply on Corent costs $0.03 to $0.10, depending on the model and the length. Typical prices for every kind of result are on the pricing page.

Do I have to pick Devstral 2512?

No. Leave the model out and Corent reads each prompt, picks the cheapest live model that does the job well (or the top one with "quality": "best"), and switches models if one fails. Pinning corent-devstral-2512 is for when you want exactly this model on every request.

More text models

Related models

Start

Run Devstral 2512 in your next request

One key opens the whole catalog: pin Devstral 2512, or leave the model out and let Corent pick. Start with $5 in prepaid credits, no subscription.