Mistral AI?!?

MR.Spuf

Elite Member
Jr. Executive VIP
Jr. VIP
Joined
May 29, 2018
Messages
6,337
Reaction score
7,054
Did anyone using: https://chat.mistral.ai/

I haven't till today, it's French (yeah yeah I know, Europe is against AI -maybe has to do with the grasp of control?!?, except France)

This is not that accurate as DeepSeek R1 or Latest o3-mini OpenAI but it's fast af. Haven't seen anything that fast yet. Faster than API!
 
Did anyone using: https://chat.mistral.ai/

I haven't till today, it's French (yeah yeah I know, Europe is against AI -maybe has to do with the grasp of control?!?, except France)

This is not that accurate as DeepSeek R1 or Latest o3-mini OpenAI but it's fast af. Haven't seen anything that fast yet. Faster than API!

It maybe having far less users
Deepseek is really good at the moment
 
I had fun testing it and comparing the results with ChatGPT. It's not bad.
I'm waiting to see what the pricing policy will be, especially since, given my use of ChatGPT, I need something that will hold up.
 
I had fun testing it and comparing the results with ChatGPT. It's not bad.
I'm waiting to see what the pricing policy will be, especially since, given my use of ChatGPT, I need something that will hold up.
Exactly, is not bad. And as I see DeepSeek API top up is unavailable atm (maybe they're increasing the costs due to their increasing base of users) so it might be a viable solution.
 
  • Like
Reactions: RK
Exactly, is not bad. And as I see DeepSeek API top up is unavailable atm (maybe they're increasing the costs due to their increasing base of users) so it might be a viable solution.
Mouais, I'm suspicious...
 
Yeah. I think deepseek got a little bit more hype than it deserves.

My opinion.
Deepseek deserves the hype. It accelerates AI in a way that every small company or even country is able to develop their AI verticals with little to no cost.
It is one of the most impactful AI breakthroughs, IMO.
The quality/benchmark of the model is not as good as expected (as the hype, but it is still very good). But as a reasoning model, it does get the job done with the correct answer and reasoning.
So it is a fantastic model all in all.
 
Last edited:
In all honesty...its better
For a tool coming out free
I agree the R1 model is very powerfull. But if you test o3-mini on an environment like hugginface, is a beast. Amd ftm DeepSeek API top up is disabled.
 
It's a good one...one more to add to our portfolio of AIs, with several tools like DeepSeek, ChatGPT, Claude, Mistral, Gemini and Perplexity, I guess we don't need to pay for a specific one and rely on it
 
It's a good one...one more to add to our portfolio of AIs, with several tools like DeepSeek, ChatGPT, Claude, Mistral, Gemini and Perplexity, I guess we don't need to pay for a specific one and rely on it
Exactly, I don't like to be super tied to just one service
 
Back
Top