LLVM's JIT support has changed a lot, but one thing didn't change: "LLVM code generation and optimization is good but expensive."
Even LLVM's "fast" code generator is slower than most JITs' code generators.
Even LLVM's "fast" code generator is slower than most JITs' code generators.
I think you can look at the people who are interested in adding LLVM tiers to their JITs (ex Apple, Facebook) to see that we're not the only ones that think there can be a place for "good but expensive" code generation in a JIT.