Director of Engineering @ Introbot.ai, scaling both code and teams with over 10 years of deep experience in systems design, performance tuning & engineering.
I travelled from Mysuru to Mumbai only to see a bunch of plastic bottles floating in the Arabian Sea 😔
Taj hotel doesn't care, none of BMC care.
I tried putting on MyBMC and they made it complicated so they can get complaints 🙄
@mybmc@CPMumbaiPolice
Xiaomi MiMo-V2.5-Pro is still the best coding model I’ve personally used.
If you can afford the pricing, it’s incredible.
Strong reasoning, reliable tool usage, fewer weird mistakes, and generally feels “senior engineer” level compared to most models.
Release dates make this even crazier:
• MiMo-V2-Flash → Dec 14, 2025
• MiMo-V2.5-Pro → Apr 22, 2026
• DeepSeek V4 Flash → Apr 24, 2026
• Gemini 3.1 Flash Lite → May 7, 2026
Competition is moving at insane speed right now.
My current takeaway:
MiMo-V2.5-Pro → best overall if budget doesn’t matter
DeepSeek V4 Flash → best value in coding/tool-calling
Gemini Flash Lite → good planner/chatty assistant
MiMo-V2-Flash → cheap flexible helper.
Thank you 😊
Yup, notifications infrastructure is working @just_snabbit 👍
At Introbot, we actually replace the core sending code with a dummy, to check the system thoroughly while not sending real messages.
Let me know if you need help 😉
The one thing I would say is necessary to have a good relationship with someone, personal or business, is mutual respect.
Like, I would not want to do anything with you if you look down at me and talk trash about me😌
Kimi K2.6, Qwen 3.6 Plus, Mimo V2 Pro look great in demos…
But in real usage? MiniMax M2.7 just feels more reliable.
Feels like most people praising the others haven’t actually shipped with them.