Mistral AI

Verified

Mistral AI is a France-based provider in the Model builder category with 24 listed models. 4 of 24 models publish a cached-input rate.

EU-owned EU hosting Model builder

Visit website API docs

At a glance

Record

Ownership: Mistral AI SAS
Headquarters: Paris, France
Hosting region: EU (Paris, Frankfurt)
Founded: Not published
Website: mistral.ai

Summary

Scope

Mistral AI is a France-based provider in the Model builder category with 24 listed models. 4 of 24 models publish a cached-input rate.

API base api.mistral.ai/v1, OpenAI-compatible. Serverless catalogue now includes third-party open weights (Z.ai GLM 5.3 / 5.3 Flash, DeepSeek V4.1 Flash 1017) alongside Mistral's own models. Batch 50% off. OCR per 1,000 pages, audio per minute, TTS per 1M chars.

Source: mistral.ai ↗ · Updated Nov 4, 2026

API access

OpenAI-compatible

Access

API

KYC

Not published

Compliance

GDPREU AI ActDPA

Key models

24 listed

Z.ai GLM 5.3 Flash 1M context · per 1M tokens Inference $0.15 / $0.5

DeepSeek V4.1 Flash 1017 1M context · per 1M tokens Multimodal $0.3 / $1.2

Z.ai GLM 5.3 1M context · per 1M tokens Inference $1.4 / $4.4

Mistral Small 4 256K context · per 1M tokens Inference $0.15 / $0.6

Mistral Large 4 1M context · per 1M tokens Multimodal $0.7 / $2.1

Ministral 3 14B 256K context · per 1M tokens Multimodal $0.2 / $0.2

Ministral 3 8B 256K context · per 1M tokens Multimodal $0.15 / $0.15

Ministral 3 3B 256K context · per 1M tokens Multimodal $0.1 / $0.1

OCR 4.1 , context · per 1k pages OCR $4 / -

OCR 3 , context · per 1k pages OCR $2 / -

Voxtral Mini Transcribe 3 32K context · per minute Speech $0.0035 / -

Voxtral Mini Transcribe Realtime 3 , context · per minute Speech $0.008 / -

Voxtral Small Transcribe 3 32K context · per minute Speech $0.004 / -

Voxtral TTS 3 , context · per 1M chars Speech $13 / -

Voxtral Small 32K context · per 1M tokens Audio $0.1 / $0.4

Codestral 25.08 128K context · per 1M tokens Inference $0.3 / $0.9

Codestral Embed 8K context · per 1M tokens Embedding $0.15 / -

Mistral Embed 8K context · per 1M tokens Embedding $0.1 / -

Mistral Moderation 2 128K context · per 1M tokens Moderation $0 / $0

Shieldstral 1.0 32K context · per 1M tokens Moderation - / -

Timestral 1.0 209M 32K context · per 1M tokens Time series - / -

Timestral 1.0 123M 32K context · per 1M tokens Time series - / -

Timestral 1.0 28M 32K context · per 1M tokens Time series - / -

Timestral 1.0 5M 32K context · per 1M tokens Time series - / -

Pricing

List rates · USD

Model Type Input price Caching price Output price

Z.ai GLM 5.3 Flash1M context · per 1M tokens Inference $0.15 $0.03 $0.5

DeepSeek V4.1 Flash 10171M context · per 1M tokens Multimodal $0.3 $0.007 $1.2

Z.ai GLM 5.31M context · per 1M tokens Inference $1.4 $0.14 $4.4

Mistral Small 4256K context · per 1M tokens Inference $0.15 - $0.6

Mistral Large 41M context · per 1M tokens Multimodal $0.7 $0.07 $2.1

Ministral 3 14B256K context · per 1M tokens Multimodal $0.2 - $0.2

Ministral 3 8B256K context · per 1M tokens Multimodal $0.15 - $0.15

Ministral 3 3B256K context · per 1M tokens Multimodal $0.1 - $0.1

OCR 4.1, context · per 1k pages OCR $4 - -

OCR 3, context · per 1k pages OCR $2 - -

Voxtral Mini Transcribe 332K context · per minute Speech $0.0035 - -

Voxtral Mini Transcribe Realtime 3, context · per minute Speech $0.008 - -

Voxtral Small Transcribe 332K context · per minute Speech $0.004 - -

Voxtral TTS 3, context · per 1M chars Speech $13 - -

Voxtral Small32K context · per 1M tokens Audio $0.1 - $0.4

Codestral 25.08128K context · per 1M tokens Inference $0.3 - $0.9

Codestral Embed8K context · per 1M tokens Embedding $0.15 - -

Mistral Embed8K context · per 1M tokens Embedding $0.1 - -

Mistral Moderation 2128K context · per 1M tokens Moderation $0 - $0

Shieldstral 1.032K context · per 1M tokens Moderation - - -

Timestral 1.0 209M32K context · per 1M tokens Time series - - -

Timestral 1.0 123M32K context · per 1M tokens Time series - - -

Timestral 1.0 28M32K context · per 1M tokens Time series - - -

Timestral 1.0 5M32K context · per 1M tokens Time series - - -

List rates. Cached input shown where published; "-" means the provider does not publish a cache rate. Units follow the source rate card.

Infrastructure

Hosting

RegionEU

CitiesParis, Frankfurt

Provider groupModel builder

ZDRnot_default (enterprise/private deployment option)

DPAAvailable

Cached input policyCached input is published per model on the docs pricing table (GLM 5.3, GLM 5.3 Flash, DeepSeek V4.1 Flash 1017, Mistral Large 4). Small/Ministral/Codestral have no cached rate.

Access

API & payments

API shapeOpenAI-compatible

AccessAPI

KYCNot published

Cached input policyCached input is published per model on the docs pricing table (GLM 5.3, GLM 5.3 Flash, DeepSeek V4.1 Flash 1017, Mistral Large 4). Small/Ministral/Codestral have no cached rate.

Compliance

Source-backed

DPAAvailable

ZDRnot_default (enterprise/private deployment option)

RetentionZDR: not_default (enterprise/private deployment option). DPA published.

Certifications

GDPREU AI ActDPA

Benchmarks

Coming soon

Coming soon.

Performance

Coming soon

Latency

Time to first token

Median and p95 response times for short and long prompts, measured from repeatable probes rather than vendor claims.

Coming soon

Throughput

Tokens per second

Streaming output rate and request throughput under steady load, with concurrency and context length recorded.

Coming soon

Efficiency

Cost per workload

Blended cost for a fixed workload, including cache reads and output tokens, so speed and price can be compared together.

Coming soon

Independent performance probes are being added. Vendor-reported numbers are not mixed into this section.