La Plateforme (Mistral API)

Offering overview

Developers reach the lab's frontier and open-weight models through La Plateforme, a hosted API that requires only payment activation to issue keys. The current lineup includes Mistral Large 4, Medium 3.5 and Small 4, the compact Ministral 3 family (3B, 8B and 14B), Codestral for code, Voxtral speech models, OCR 4.1, Mistral Embed and Codestral Embed, plus moderation models, with a third-party GLM model also served. Many weights are Apache 2.0 licensed, so the same models can be self-deployed with vLLM or TensorRT-LLM. Mistral Studio layers agents, persistent workflows, enterprise connectors, evaluation tooling and observability on top of the API. Enterprise customers can opt for hybrid, dedicated or fully self-hosted deployments to keep data inside their own perimeter, a selling point for European buyers seeking an alternative to US providers. Billing is per million tokens with input and output counted separately; batch jobs halve the rate and cached input tokens are discounted up to 90 percent. The public pricing page quotes Mistral Large at $0.50 per 1M input tokens and $1.50 per 1M output tokens, with smaller models positioned for cost-sensitive work.

Technical specifications

Containers

Unknown

Kubernetes

Unknown

Root access

Unknown

API access

Unknown

Autoscaling

Unknown

Persistent processes

Unknown

AI capabilities

Model serving

Mistral Large 4, Mistral Medium 3.5, Mistral Small 4, Mistral Large 3, Ministral 3, Codestral, Codestral Embed, Mistral Embed, Voxtral, Mistral OCR 4.1, Shieldstral, Mistral Moderation 2, GLM 5.3

AI frameworks

Not stated

Inference endpoints

Unknown

Fine-tuning

Unknown

Vector database

Unknown

Model catalog

Unknown

Monitoring

Not stated

Evaluation

Not stated

Security & compliance

Geography

Provider headquarters are separate from offering infrastructure and availability.

Available plans

Pricing can change. Promotional prices are shown separately from standard prices.

Questions & answers

Which models can developers access through La Plateforme?

The lineup includes Mistral Large, Medium and Small, Ministral, Codestral, Voxtral, OCR, Mistral Embed, Codestral Embed and moderation models. A third-party GLM model is also served.

What is required to issue API keys?

The hosted API requires payment activation to issue keys.

What does Mistral Studio add to the API?

Mistral Studio adds agents, persistent workflows, enterprise connectors, evaluation tooling and observability.

What deployment options are documented?

Enterprise customers can choose hybrid, dedicated or fully self-hosted deployments. Many model weights are Apache 2.0 licensed and can be self-deployed with vLLM or TensorRT-LLM.