$100M-500M gets thrown around [1] but it's hard to find good data. This might seem not too high, but many drugs fail, and often fail at the most costly step (phase III). So you have to be sure to factor that failure rate in.
Depends on what indication they are trying to get. A handy estimate is that a phase III trial costs $15K/person/year. If they aim for a 1000 person trial of 6 month duration, you're looking at $15M just for the phase 3 (I doubled it since they typically run two). However, that's a small trial.
Fidaxomycin, which was developed by Optimer Pharmaceuticals to treat C. difficle, was relatively inexpensive to run. I want to say in the vicinity of $100 million
