standard errors from glmfit and fitglm are different

The screenshot is attached.

I used the two commands below. I think the se of the two mothods should be the same. No idea about the difference.

I appreciate your time.

[b,dev,stats] = glmfit(num(:,[2:4,9:12]), num(:,T), 'poisson','estdisp','on','offset',num(:,5));

m = fitglm(num(:,[2:4,9:12]), num(:,T), 'Distribution','poisson','offset',num(:,5))

note: num is my data table.

[WEN XIE]

And is there any one who knows how to do negative binomial regression in matlab? I set the dispersion parameter of poisson model. Is that the same as negative binomial model? Thank you.