OBJECTIVES: To investigate the reliability of bi-axial flexure and three-point flexure testing of an experimental, low-shrink oxirane-based dental composite, EXL596 compared with two conventional methacrylate based restoratives, Z250 and Z100. METHODS: Specimens (n=20) of a novel oxirane-based composite and two commercially available methacrylate based composites were fabricated for flexural testing to evaluate mean bi-axial flexure strengths, three-point flexure strengths and the associated Weibull moduli (m) following 24 h immersion in a lightproof waterbath maintained at 37+/-1 degrees C. RESULTS: Mean bi-axial flexure strengths and the associated m of EXL596, Z250 and Z100 were 168+/-11 MPa (m=16.2+/-4), 140+/-12 MPa (m=11.9+/-3) and 126+/-13 MPa (m=10.2+/-2), respectively. Three-point flexure strengths and the associated m of EXL596, Z250 and Z100 were 113+/-15 MPa (m=9.2+/-2), 92+/-10 MPa (m=8.5+/-2) and 79+/-16 MPa (m=6.3+/-1), respectively. CONCLUSIONS: The present investigation suggests that bi-axial flexure strength testing of dental resin-based composites provides a more reliable testing method than three-point flexure. The increased reliability was considered in terms of the associated Weibull moduli following bi-axial flexure testing as a result of the elimination of the additional induced variability introduced during the curing regime of three-point flexure specimens.