Organization ProfileMistral AI SAS

About Mistral AI

Founded in 2023 in Paris by former researchers from Meta and Google DeepMind, Mistral AI is Europe's premier foundation model company. Known for pioneering sparse Mixture-of-Experts architectures with Mixtral 8x7B, Mistral provides efficient models on Bedrock that emphasize high throughput, code comprehension, and multilingual fluency.

HeadquartersParis, France
Founded2023
OwnershipPrivately held
Flagship LineageMistral Large, Mistral Small, Mixtral (8x7B, 8x22B), Codestral
Model Ecosystem & LicensingDual approach: Apache 2.0 open weights & commercial endpoints

Mistral AI Token Cost Estimator

Calculate estimated monthly inference spend for Mistral AI foundation models.

Estimated Monthly Spend
$0.00
Prompt (Input):$0.00
Prompt Cache Read:$0.00
Completion (Output):$0.00
—

All Mistral AI Models (14)

Model NameTierModalityInput $/1MOutput $/1MBatch In $/1MBatch Out $/1MCache Read $/1MRegions
DevstralMistral
GeneralText$0.40$2.00$0.20$1.00—13
GeneralText$0.50$1.50$0.25$0.75—13
LightweightText$0.20$0.20$0.10$0.10—13
LightweightText$0.10$0.10$0.05$0.05—13
LightweightText$0.15$0.15$0.07$0.07—13
Mistral 7BMistral
LightweightText$0.15$0.20———9
FlagshipText$4.00$12.00———9
FlagshipText$2.00$6.00$1.50$4.50—1
FlagshipText$0.50$1.50$0.25$0.75—13
GeneralText$1.00$3.00$0.50$1.50—1
LightweightText$0.45$0.70———9
FlagshipMultimodal$2.00$6.00———7
LightweightAudio / Speech$0.04$0.04$0.02$0.02—13
GeneralAudio / Speech$0.10$0.30$0.05$0.15—13