Update deepseek-r1-distill-llama-8b.mdx

Author: fpagny

pages/managed-inference/reference-content/deepseek-r1-distill-llama-8b.mdx

@@ -19,23 +19,21 @@ categories:
 |-----------------|------------------------------------|
 | Provider        | [Deepseek](https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Llama-8B)  |
 | License        | [MIT](https://huggingface.co/datasets/choosealicense/licenses/blob/main/markdown/mit.md)  |
-| Compatible Instances | L4, H100, H100-2 (FP8, BF16) |
-| Context Length | up to 32k tokens |
+| Compatible Instances | L4, H100 (BF16) |
+| Context Length | up to 131k tokens |
 
 ## Model names
 
 ```bash
-meta/deepseek-r1-distill-llama-8b:fp8
-meta/deepseek-r1-distill-llama-8b:bf16
+deepseek/deepseek-r1-distill-llama-8b:bf16
 ```
 
 ## Compatible Instances
 
 | Instance type  | Max context length |
 | ------------- |-------------|
-| L4      | 32k (FP8, BF16) | 
-| H100      | 32k (FP8, BF16) |
-| H100-2      | 32k (FP8, BF16) |
+| L4      | 39k (BF16) | 
+| H100      | 131k (BF16) |
 
 ## Model introduction
 
@@ -47,7 +45,7 @@ DeepSeek R1 Distill Llama 8B is designed to improve performance of Llama models
 It is great to see Deepseek improving open(weight) models, and we are excited to fully support their mission with integration in the Scaleway ecosystem.
 
 - DeepSeek-R1-Distill-Llama was optimized to reach accuracy close to Deepseek-R1 in tasks like mathematics and coding, while keeping inference costs limited and tokens speed efficient. 
-- DeepSeek-R1-Distill-Llama supports a context window up to 32K tokens and tool calling, keeping interaction with other components possible.
+- DeepSeek-R1-Distill-Llama supports a context window up to 131K tokens and tool calling, keeping interaction with other components possible.
 
 ## How to use it