Polyetheretherketone (PEEK) is used as an alternative to titanium in medical devices. Previous in vitro studies examining PEEK have differed in their choice of polymer variant [PEEK or carbon-fiber reinforced PEEK (CFR-PEEK)], source of polymer (some of which are no longer available or for implantation) and cell type. While all studies demonstrated favorable cytocompatibility of the PEEK material, no studies are available which reflect the current state of the art of the material. Here, we use different forms of the only implantable grade PEEK available. These are compared with commercially pure titanium (cpTi) Grade 1 using a human primary osteoblast model. Sample materials were presented as industrially relevant surfaces. Machined or injection molded PEEK and CFR-PEEK were evaluated along with polished (Ra=0.200microm) and rough (Ra=0.554microm) cpTi. Osteoblast adhesion at 4h on injection molded variants of PEEK (Ra=0.095microm) and CFR-PEEK (Ra=0.350microm) material was comparable to titanium. Machined variants of PEEK (Ra=0.902microm) and CFR-PEEK (Ra=1.106microm) materials were significantly less. Proliferation at 48h determined by [(3)H]-thymidine incorporation was the greatest on the smoothest of all materials, the injection molded unfilled PEEK, which was significantly higher than the rough titanium control. The machined unfilled PEEK had the lowest DNA synthesis. RT-PCR for alkaline phosphatase, Type I collagen and osteocalcin normalized to glyceraldehyde-3-phosphate dehydrogenase revealed different patterns of mRNA levels. High mRNA levels for Type I collagen showed that CFR-PEEK stimulated osteoblast differentiation, whilst injection molded unfilled PEEK was less differentiated. Machined unfilled PEEK had comparable message levels of bone matrix proteins as rough titanium. All material variants permitted a degree of mineralization. Scanning electron microscopy at 3 days and 2 weeks in differentiation medium showed that human osteoblasts were well spread on all the different substrates. The varied response reported here at different time points during the study suggests that material formulation (unfilled PEEK or CFR-PEEK), subjection to industrial processing, surface roughness and topography may all influence the cellular response of osteoblasts to PEEK. Thus, differences in human osteoblast responses were found to the various samples of PEEK, but implantable grade PEEK, in general, was comparable in vitro to the bone forming capacity of rough titanium.