What scientific term would contain the greatest amount of information that scientists consider to be true to the best of their knowledge?
A few thoughts...
If we are talking about natural sciences, then it seems to me that any scientific hypothesis, conjecture or theory that asserts the truth of a model is by nature unprovable.
We can demonstrate that observations are consistent (or inconsistent) with our models, but the notion of our models being true lies beyond what science can do.
So the scope of "true to the best of their knowledge" is somewhat limited.
The sort of hypothesis that we may consider proved beyond reasonable doubt may be something like "The Earth is more than 4.5 billion years old". Even such a statement may be challenged by what we might consider unreasonable hypotheses.
If we are talking about the so-called pure sciences, then various statements may be provably true, but only within their abstract contexts. When we apply abstract models to reality, we should recognise that that is what we are doing.