Vertex AI

“I've always had to go through Vertex or OpenRouter for any sort of heavy load”

“Vertex Cloud needed 2 years to get OAI/Ant style API keys working and were losing dev mindshare”

“there is vertex api and rebranded to Gemini something and making it very confusing”

“Their Zero Data Retention was in fact contractually enforced”

“claims that "Claude 3 Opus" [sic] was a state of the art model we could use from vertex.”

“the UX is a lot worse, and it can be hard to estimate cost since it's typically lagging”

“With all the encrypted at rest stuff and other guarantees that vertex provides.”

“the main provider i know that just doesn't go down is vertex and they charge 2-3x the rest”

“the same models are hosted by both AWS and Google with better uptime than Anthropic”

“serves up other companies models better than they can themselves”

“Likely, and via vertex on gcp (or whatever they are calling it this year).”

“routing exclusively through Vertex AI for enterprise security”

“Every time I try to decide whether to use Vertex AI or something else, I end up confused”

“I've found the Vertex API seems to be significantly less strict.”

“When I run production workloads on vertex or bedrock i am routinely confronted with quotas”

“quota for Opus models defaults to 0 on VertexAI and quota increase requests are auto-rejected”

“You can access Claude models with Google Cloud reliability via VertexAI.”

“Higher limits and quota, goes up with more spend”

“They don't read the things you send them, not even for "safety checks" or sys-admins accessing the system.”

“A privacy policy that's at least as good as Vertex.ai at Google.”

“much better, includes Claude models and many of the open weight models too”

“use it over Vertex AI, which allows you to generate anyone”

“they are separate from your personal (internet browsing) identity and have clear language that your inputs and interactions are not used for training”

“it'll probably be a better idea to run the models hosted in Amazon Bedrock or Google Vertex”

“Have not personally had as big an issue figuring out Vertex.”

“which uses vertex ai and all models therein”

“I pay by the token via VertexAi, better than those budget plans that rate limit you”

“Or tried not to get randomly ratelimited with a single request to a vertex llm model?”

“I don't see anything you do that they don't already do for me”

“Anthropic and Gemini models have been available on Vertex without relying on GPUs for a while.”

“consider this for addtl cost savings if local doesnt interest you”

“until pretty recently using that took figuring out service accounts”

“much more stable but lacking in some functionality”

“Now there's "express mode" for Vertex which uses an API Key, so things are better”

“I'm using it fine through both aistudio and vertex ai, direct API calls”

“Google was nice enough to launch Vertex AI inside GCP for enterprises”

“If you're already using GCP, Vertex AI is pretty good.”

“I've never had the downtime or service busy situations I've heard others complain about with other vendors.”

“if GOOGLE_GENAI_USE_VERTEXAI=true in your environment, woe betide you if you're trying to use gemini cli with an API key.”

“The documentation is fragmented with redundant offerings making it confusing to determine what is what.”