diff --git a/docs/api-reference/models.md b/docs/api-reference/models.md
index 45ae29f..ac76c74 100644
--- a/docs/api-reference/models.md
+++ b/docs/api-reference/models.md
@@ -259,7 +259,7 @@ K3 is the current flagship — a 2.8-trillion-parameter open MoE with a **1M-tok
 
 | Model ID | Name | Input Price | Output Price | Context |
 |----------|------|-------------|--------------|---------|
-| `mistral/mistral-large-4` | Mistral Large 4 (reasoning, image input) | $0.68/M | $2.09/M | 512K |
+| `mistral/mistral-large-4` | Mistral Large 4 (reasoning, image input) | $1.36/M | $4.18/M | 512K |
 
 ### Xiaomi
 
diff --git a/docs/products/intelligence/pricing.md b/docs/products/intelligence/pricing.md
index 1118dc0..951e684 100644
--- a/docs/products/intelligence/pricing.md
+++ b/docs/products/intelligence/pricing.md
@@ -151,7 +151,7 @@ Grok doubles the per-token rates above 200K prompt tokens (the whole request rep
 
 | Model | Input (per 1M) | Output (per 1M) | Cached input (per 1M) |
 |-------|----------------|-----------------|-----------------------|
-| Mistral Large 4 | $0.68 | $2.09 | — |
+| Mistral Large 4 | $1.36 | $4.18 | — |
 
 ### DeepSeek