diff --git a/snapshots/mistral/models.md b/snapshots/mistral/models.md
index 30dcb26..2f39ee5 100644
--- a/snapshots/mistral/models.md
+++ b/snapshots/mistral/models.md
@@ -5,49 +5,55 @@
 Mistral develops open-weight and commercial large language models. Explore the full lineup, compare benchmarks, and find the right model for your use case.
 
-Latest models
+Explore
 
-### Mistral Medium 3.5
+Compare specifications and benchmarks, find the right model for your task, and learn how to deploy across cloud providers.
 
-Our frontier-class multimodal model optimized for agentic and coding use cases.
+### Overview
 
-v3.5
+View all models with specifications, benchmarks, and feature comparison.
 
-### OCR 3
+### Model selection guide
 
-Our OCR service powering our Document AI stack.
+Choose the right model based on task requirements, latency constraints, and cost targets.
 
-v25.12
+### Deployment
 
-### Voxtral Mini Transcribe Realtime
+Deploy models to cloud providers and Mistral Compute.
 
-An efficient audio input model, fine-tuned and optimized for live transcription purposes.
+### Best practices
 
-v26.02
+Prompt engineering, sampling parameter tuning, and evaluation methods.
 
-### Devstral 2
+Latest models
 
-Our frontier code agents model for solving software engineering tasks.
+### Mistral Medium 3.5
 
-v25.12
+Our frontier-class multimodal model optimized for agentic and coding use cases.
 
-View all models
+v26.04
 
-Explore
+### Mistral Small 4
 
-Compare specifications and benchmarks, find the right model for your task, and learn how to deploy across cloud providers.
+Hybrid model unifying instruct, reasoning, and coding in a single efficient model.
 
-### Overview
+v26.03
 
-View all models with specifications, benchmarks, and feature comparison.
+### Voxtral Mini Transcribe Realtime
 
-### Model selection guide
+An efficient audio input model, fine-tuned and optimized for live transcription purposes.
 
-Choose the right model based on task requirements, latency constraints, and cost targets.
+v26.02
 
-### Deployment
+### Voxtral TTS
 
-Deploy models to cloud providers and Mistral Compute.
+State-of-the-art TTS with zero-shot voice cloning and multilingual support.
 
-### Best practices
+v26.03
+
+### OCR 3
 
-Prompt engineering, sampling parameter tuning, evaluation methods, and fine-tuning workflows.
+Our OCR service powering our Document AI stack.
+
+v25.12
+
+View all models