I've been using the Deep Research feature of ChatGPT, which was included in the Plus subscription. This feature was really helpful because it provided reliable scientific references instead of just generating text without backing. However, it's now been moved to a higher subscription tier at ten times the cost. I get that it requires a lot of processing power, but I'm exploring options. I've managed to get DeepSeek running locally, but I'm wondering if there's a way to replicate this Deep Research feature with DeepSeek or other models. Any suggestions for alternatives?
7 Answers
You can also find this type of feature available for free on Gemini and Grok, so those are worth checking out.
I'm developing one using npcpy. Would you be interested in collaborating on it?
You might want to check out Qwen3; it has a similar option, although the quality isn't quite up to OpenAI's level. I'm hoping DeepSeek eventually adds something comparable, but right now, they're focusing more on their core models than on new tools.
Claude has been performing well for research tasks. It does require a bit more back-and-forth to get the best results, but it's not under a subscription like OpenAI's now.
I'm a fan of Deep Research too, but I have to limit my usage on my ChatGPT Teams account and I don't want to switch to pro. I recently got a Gemini advanced account, and I can do up to 20 queries a day, which has worked out well for me, only hitting my limit once.
There are a couple of projects on GitHub that could help you get DeepSeek to mimic that feature. Perplexity is also worth looking into; it can provide some similar functionalities.
By the way, you can find it in that dot bar, the bar dot button. Here’s a preview to help you locate it!
Here's another image I found to help too!
Why recommend Qwen when you could go for Gemini? Seems like you're pushing those Chinese models too hard.