I'll rank them in order of how much I like them as well as criticisms I have
1) Kardar - Statistical Physics of Particles. It's a good solid book, succinct and to the point. Doesn't cover anything but core topics. Generally the exposition is pretty good and I have no real quibble about it. My main gripe is how hard his problems can be, and the amount of knowledge you have to sometimes bring from outside sources. Fortunately a lot of solutions are in the back, some detailed some not.
2) Chandler - Modern Statistical Mechanics. Similar to Kardar but lacks some topics that I find to be core, most notably the BBGKY hierarchy (which is treated in Kardar!). I feel like he could have done away with his chapter on monte carlo numerics and added better coverage but whatever. Up side, with some inventive googling you can find the solutions manual.
3) Huang - Statistical Mechanics. Some people like him, some don't. I think his book is just fine. Good coverage of core topics, and some extra topics like the Darwin-Fowler method are pretty nice. Problems are of varying difficulty. He has a set of chapters dealing with special topics which I wouldn't spend much time on, except his chapter on the Onsager solution of the 2d Ising model. Most books only present the Vdovichenko solution (a combinatorial solution) as first popularized in the west by Landau and Lifgarbagez. Otherwise, there are better books to learn renormalization and critical phenomena from.
The rest of the books I'll list are what I would only use for supplementary reading, if given the choice.
4) Reichl 2nd ed. The second edition covers a lot of topics, a lot of them aren't core. She includes a fair amount of example problems, which is nice, however the book is entirely too verbose. Similarly, the 3rd edition had almost half its pages removed and accelerates too quickly from core to specialized topics.
5) Pathria. I didn't like his book. Skips steps and quotes results too much for my tastes. Very verbose for saying little. His book is used a lot in some courses, but I think that's changing. Kardar is growing in popularity.
6) Landau and Lifgarbagez. Way too much coverage of topics. As good as the rest of the books in the series, but I wouldn't consider it primary learning material.
7) LeBellac. I used this book for my stat mech course. Ultimately I didn't like it, but my teacher was fantastic so I have some very detailed notes. It suffers the same problem as Pathria, a lot of results quoting and steps skipped.
If you want some graduate level problems and solutions there's Kubo, Dalvit, and Landsberg. All separate books.