With sufficient context, I find pretty much any of the modern ones to be sufficient. If you spend the time to write the prompt, they work pretty well. I use Claude at work to automate some of my more common tasks. But I find that using a basic model to help me develop python code that does the same thing is actually faster.
In the cases where I actually rely on something to solve a niche problem, all of them suck.
Yes, I usually use Claude to do some code tasks. And I didn't try other AI Agent products. I am not sure whether there's one AI agent that is very popular, but I don't know it.
With sufficient context, I find pretty much any of the modern ones to be sufficient. If you spend the time to write the prompt, they work pretty well. I use Claude at work to automate some of my more common tasks. But I find that using a basic model to help me develop python code that does the same thing is actually faster.
In the cases where I actually rely on something to solve a niche problem, all of them suck.
Yes, I usually use Claude to do some code tasks. And I didn't try other AI Agent products. I am not sure whether there's one AI agent that is very popular, but I don't know it.
I like Gemini 2.5 Pro, Claude and Perplexity Pro
1. Qwen 2. Grok 3. Gemini
Qwen is good? I haven't even tried it.