Creep behavior of Alloy 617 under multi-axial loading has been characterized at 950°C using thin walled tubes with internal gas pressure. Tubes were machined from Alloy 617 plate material to ensure comparability of material properties with extensive testing carried out in the VHTR Research and Development program. Properties of this alloy under multi-axial loading are of interest for elevated temperature design in advanced nuclear heat transfer systems. A constitutive model in the form of a Norton relationship was developed for Alloy 617 plate to allow development of a suitable creep specimen using finite element simulation. It was determined that proper description of the minimum creep rate at 750°C required a threshold stress formalism. The magnitude of the threshold stress was determined to be 75MPa for creep tests using the conditions examined. This threshold stress arises from strengthening by the γ'(Ni 3 Al,Ti) intermetallic phase; above this temperature this phase is unstable and Alloy 617 behaves as a solid solution. Comparison of strengthening models to detailed TME analysis of dislocation-particle interaction in crept specimens indicated that strengthening arises from localized climb, or a combination of climb and Orowan bowing depending on the creep time and stress. Quantification of the strengthening from γ' is an important result to confirm the region over which it is acceptable to extrapolate creep rupture times on a Larson-Miller plot. This result will also help inform the activities underway in the US and Europe to develop a γ' strengthened version of Alloy 617 for application in ultra-supercritical fossil power plants. Failure of the pressurized tubes initiates at low strain by cracking of the majority of grain boundaries that intersect the exterior surface. For radial strains of about 5% or less the total creep damage fraction is equivalent to uniaxial test specimens with similar tertiary creep strain. Above this value the surface nucleated cracks begin to open rapidly with increasing strain and the volume fraction of creep damage exceeds that for comparable uniaxial tests. Final failure, defined for the pressurized tubes as loss of the ability to maintain pressure, occurs by linking of damage ahead of one of the cracks that nucleate on the surface on a 45° angle to the tube radius. Although the time to tube rupture was comparable to uniaxial creep rupture on the Larson-Miller diagram, cavitation was more extensive and the creep strain was considerably reduced to plate. The minimum creep rate determined from pressurized tube results appears to fit a different Monkman-Grant relationship compared to uniaxial creep data.