client AI context

Anthropic recently released Claude 4.5 Opus. I wanna talk about what we're using it for and also where we're still using chat, t, Gemini or even Claude 4.5 Sonic.

So as I've spoken about in other videos, we regularly use chat, G, B, Gemini and Claude. However, Claude is the one that we use most often for work.

has always been the best at considering a large amount of context without forgetting details and the best at natural writing, which is our primary use case for large language models. And so we set up a Claude project for each of our clients and use that to build their brand ambassador.

This is just a project that we give all the information there is about a particular client so it can represent them. One common problem with Claude though is that it frequently runs out of tokens.

This is because when you give Claude a large amount of context, it really pays attention. It'll consider all that context where other models are quicker to switch over to rag and only pull in the information it thinks is important to the prompt.

This has always been the catch 22 with Claude...

strategy or analysis or organization, Opus is the best choice. On the other hand, if we want the model to help us retrieve the context, if we need the model to do any kind of search or research across the internet, Gemini 3 is definitely the better choice.

And for natural writing, if we're writing a blog post, social media post, webpage copy, sonic 4.5 still seems like the best option. So where does that leave ChatGPT?

I'm not sure yet. We're actually switching most things that we used to use ChatGPT for over to Gemini 3 right now.

I will say that ChatGPT's Thinking mode and Deep Research mode does try a bit harder than Gemini's, just not quite as smart. And ChatGPT search function is much less sophisticated than Geminis.

I do know that ChatGPT Codex is very powerful. It's just not something that we use in our work.