i understand you want people to test the new Gemini pro thinking model But why remove the normal pro model ?
it’s literally the best model i have ever used.
edit: I take back my words the new amount of tokens the model can output at a time is increased to 65536!!
that is just super awesome!
10 Likes
it is not actually 65536. it is sum of thinking output tokens + output tokens.
2 Likes
Thinking Gemini 2.5 Pro is great, but it’s pretty weak in writing, editing and translations (most thinking models are, as far as I tested). And Gemini 2.0 Pro was great in exactly these areas (better than GPT 4.5). It’s a shame to not have it anymore, because it was the best translation and editing AI tool out there.
4 Likes
lol man it used to be 8000 something now its litterly a digit more that is way more than the thinking step
the model now can output 3000+ code lines
1 Like
Yeah, it feels a little insane that they’re just going to remove the 2.1 pro model without any proper replacement like Flash is clearly not intended for use with writing and proper creative thinking or brainstorming, and 2.5 might have been a decent replacement except it’s a thinking model, and it takes so long just to process any input, and it appears (from my limited use so far) to have severely lower performance on creative writing tasks
1 Like
You can still access it using vertex APIs
2 Likes
well that is actually nice to know 
1 Like