Dear,

The state of the tools are undoubtedly useful, however accurateness depends on your criteria. If you are comparing with crystallography results, we are not there yet. However if you want to have an idea about relative accurateness of the tools you can look at the results of the CASP challenges, where tools are welcome to predict a known structure using their algorithm.

As a rule of thumb smaller protein sequences will have higher probability of yielding more accurate results compared to larger (let's say 2000aa) protein sequences, that being said it does not guarantee an absolute margin over larger ones.

See this post, it wll help you.

Protein Structure tools

There are some other resources:

Improving prediction of secondary structure, local backbone angles, and solvent accessible surface area of proteins by iterative deep learning

http://www.nature.com/articles/srep11476

Customised fragments libraries for protein structure prediction based on structural class annotations

http://bmcbioinformatics.biomedcentral.com/articles/10.1186/s12859-015-0576-2

Bayesian Model of Protein Primary Sequence for Secondary Structure Prediction

http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0109832

Expasy-tools

https://www.expasy.org/proteomics/protein_structure

book-chapter:

http://zhanglab.ccmb.med.umich.edu/papers/2009_4.pdf

YASARA is a molecular-graphics, -modeling and -simulation program for Windows, Linux, Mac OS X

http://www.yasara.org/

Concerning the site you have mentioned:

http://zhanglab.ccmb.med.umich.edu/I-TASSER/