Climate, while much simpler than weather, is still rather more complicated than school grades.
So 100% accurate ? No. But not 50% either - more in the 95%+ range. That's also not an entirely true assessment because you're measuring it the wrong way. Climate models are written by experts who are aware they can't factor in everything, and that some things are still being worked on, so they don't give you an exact temperature - they give you a range within which the outcome is likely to lie and, if you take the average over the period predicted for, they overwhelmingly do lie in those averages.
The one major discrepency is IPCC reports, there are several decades where the average warming was significantly higher than IPCC models predicted. The reason for this is that the IPCC is particularly conservative in their estimates, fear of being called alarmists have led to the IPCC only publishing the bottom end of the likely range and also excluding anything they don't have extremely high confidence in (far higher than any other science would need for a minor variable in a big set with limited influence) - as a result they tend to to somewhat under-predict warming.
The lesson from that is that IPCC reports should be read as an absolute best-case scenario, reading the papers they are based on - the upper limit worst-case scenarios should be considered as well and we can generally expect reality to lie somewhere in the middle between those.,
It's posts like yours that turn people into skeptics.
1. You want to claim that climate models as a whole are 95% accurate.
2. You want to claim there are several decades where warming exceeds the IPCC prediction because they are so conservative and are clearly biased to a best-case scenario.
1. On what basis can you claim 95% accuracy for models, because it surely isn't in a predictive sense. Climate models have NOT managed to predict future climate with 95% accuracy. More accurately, if I said temperature averages for the next 30 years will be similar to this year +/- 0.5c, I'd have great accuracy, but I really haven't predicted a bloody thing either, now have I? The real benchmark used for climate modelling is still exclusively hindcasts, which means nobody is successfully publishing models that fail to hindcast accurately, which proves exactly nothing about the accuracy of the modelling. What is more, the basic physics are pretty well understood, the global energy imbalance at the edge of the atmosphere is the problem that the greenhouse effect is driving. Climate models regrettably still require manual fine tuning to poorly understood variables like clouds to prevent unrealistic runaway energy imbalances. So I don't think your characterisation of climate models in anyway reflects reality.
2. The first IPCC report is 26 years old, would you care to point out which decades it grossly underestimated? There are really only 2 full decades there to pick from so I'm interested to hear how you managed to get that plural usage of decades in. Also, if you want to look at the older IPCC reports for comparisons, the real world temperatures since the Third assessment in 2001 have trended right near the bottom end of the error bars of the most optimistic IPCC projections right up until a year ago.
Seriously, we have overwhelming evidence that the planet is warming and that our CO2 emissions are contributing meaningfully to that. You don't need to make up stupid exaggerations of the known science to try and make it sound scarier than it is. That's not science. That's not defending science. It's actively undermining and misrepresenting science. False statements like yours lead people to question and doubt everything else they hear when they find out your claims were bunk.