Mistral AI

Perception

Highly efficient multimodal model for vision and language tasks.

Strong vision-language understandingEfficiencyOpen weightsGood performance on benchmarks
Today's score
90.0
Try Perception

Where it ranks today

Best for / Not great for

Best for
  • Image captioning
  • Visual question answering
  • Document analysis
  • Embedding generation for multimodal search
Not great for
  • Complex audio processing
  • Generating video content
  • Highly abstract reasoning tasks

Why it ranks here

Mistral AI's Perception model is gaining traction for its impressive vision-language capabilities combined with open-weight accessibility. Its efficiency and strong benchmark performance make it a compelling choice for developers looking for a robust, customizable multimodal solution.

30-day trend

Score breakdown

Search trends89
Benchmarks91
Developer buzz92
News mentions90

Pricing

API: $0.00 in · $0.00 out per 1M tokens · Consumer: $0.00/mo

Pricing plans

Mistral Small
Efficient model for quick tasks.
Free
  • Fast inference
  • Lower cost
  • Text generation
Try Small
Mistral Large
Top-tier performance and reasoning.
$40/mo
  • Multilingual capabilities
  • Strong reasoning
  • Web search integration
Try Large
Popular
Perception (API Access)
Powerful vision-language model.
$10/mo
  • Open weights
  • Image and text understanding
  • Efficient processing
  • Customizable API
Access Perception
Compare with another modelHow is this score calculated? →Snapshot 2026-08-01