Can Gemini 4 Hydrogen Solve Repetitive Questions? 1M Context and 95% Cache Price Reduction Are Not Persistent Memory

Written by

in

Shattered’s analysis of Gemini 4 Hydrogen points out that its 1,048,576-token context can reduce repeated input within a single task, while the 95% reduction in cached input pricing can lower the cost of resending long prefixes. However, neither mechanism amounts to memory across sessions.

Source:Shattered · Read original

Comments

Leave a Reply

Your email address will not be published. Required fields are marked *