Do Pro models such as DeepSeek V4 Pro and Kimi K3.1 Pro underperform expectations?
AI summary
Some views question whether several recent models ending in Pro have underperformed expectations, naming DeepSeek V4 Pro, Kimi K3.1 Pro, and the not-yet-released Gemini 3.5 Pro. The poster thus suggests that making Flash models is the right path for most AI teams.
Body · AI translation
Recently, haven't the models ending in Pro all been somewhat below expectations?
DeepSeek V4 Pro, Kimi K3.1 Pro...
And the stillborn Gemini 3.5 Pro...
So is making Flash models the right path for most AI teams?