max output token and knowledge cutoff for Mistral-8B-Instruct-2410

[zhoumengbo]

Python -VV

Hello,

I would like to ask two questions about the Mistral-8B-Instruct-2410:

1. **What is the maximum number of output tokens** the model can generate during inference?
   - For example, is there a known limit like 2048, 8192 tokens?

2. **What is the knowledge cutoff date** for this version?
   - Is the model trained on data up to a specific month or year (e.g., 2023-03, 2023-08, etc.)?

I’ve searched the documentation but couldn’t find a definitive answer to these two questions.

Thank you in advance for your help!

Pip Freeze

None

Reproduction Steps

None

Expected Behavior

None

Additional Context

No response

Suggested Solutions

No response

[darkb0ts]

maximum number of output tokens the model can generate during inference is not strictly fixed, but it’s commonly limited by the context window, which is 8192 tokens for Mistral-8B models.