1 Like
Waiting and watching
1 Like
At this price, unless it completely outperforms Deepseek in every way, I think it’s really expensive.
It burns through credit really fast. I just topped up Paygo with 5 bucks and asked it to help me tweak a UI. After about half an hour, the 5 bucks were all gone.
The token plan burns even faster.
I just found out that after Mimo lowered its price, it’s exactly the same as DeepSeek—neither of them supports multimodal, and with Mimo, if you upload an image at any point during your session, you have to clear everything.
Anyway, before multimodal comes to ds, I’ll keep using kimi and mimo, and just use ds as a support.
DS and MIMO are still a bit lacking compared to each other — still no multimodal capabilities, and the main issue is the price 
mimo has multimodal capabilities.
I’m burning through it way too fast!! I can’t keep this up.
There isn’t a 2.5 Pro.
2.5 has it, v2-omni has it too,
There’s no point in nitpicking over words.
Domestic models have never lost in benchmarks.
We’re actually talking about different entry points. I mostly use it within cc and Opencode, and I’ve always used 2.5pro with Mimo, which doesn’t support vision. There’s a weird issue: if you try to input or involve any images during the process, it throws an error.
Context is not compatible — multimodal and non-multimodal models can’t be mixed. When using them, don’t switch models in the middle. If you need another model to do something, open a sub-agent or a new window.