Gemini 2.0: our new AI model for the agentic era

Re:Gemini 2.0: our new AI model for the agentic era by manishsharan on Wednesday December 11, 2024 @11:14AM (#42389274)

>> hard to deny that their LLM models aren't really, really good though.

The context window of Gemini 1.5 pro is incredibly large and it retains the memory of things in the middle of the window well. It is quite a game changer for RAG applications.

[ Parent | on HN ]

Re:Gemini 2.0: our new AI model for the agentic era by KaoruAoiShiho on Wednesday December 11, 2024 @02:14PM (#42391566)

It looks like long context degraded from 1.5 to 2.0 according to the 2.0 launch benchmarks.

[ Parent | on HN ]

Re:Gemini 2.0: our new AI model for the agentic era by verdverm on Thursday December 12, 2024 @01:11AM (#42396674)

Those benchmarks are for the Flash 2.0 model, showing it improved over the 1.5 Pro model minus for that one benchmark

[ Parent | on HN ]

Re:Gemini 2.0: our new AI model for the agentic era by caeril on Wednesday December 11, 2024 @02:44PM (#42391955)

Bear in mind that a "1 million token" context window isn't actually that. You're being sold a sparse attention model, which is guaranteed to drop critical context. Google TPUs aren't running inference on a TERABYTE of fp8 query-key inputs, let alone TWO of fp16.

Google's marketing wins again, I guess.

[ Parent | on HN ]