It may depend on the ligand coordinating the centre metal ion. But how do you confirm the reduction in the cluster?

If the precursor is Co(II) and the ligand backbone is known [suppose contains  four O1-] then after solving the crystal, depending on the conter cation number the charge of the central Co can be understood.

secondly, If the central metal is Co(III) then the complex will be NMR active but EPR silent and if in Co(II) state then EPR active and NMR silent.

Thirdly, If Co(II) exists then room temperature magentic study will also help to determine the magnetic moment (SQUID is not always available).

Thanks.

Bye!

@Pritam Ghosh: The methods you proposed holds good for usual metal complexes but for cluster the scenario is more complex. In cluster, not all the Co-ions may have the similar OS and therefore neither NMR nor EPR may give you a good results. The crystal structure for cluster may give some indication but OS cannot be directly determined by it. A magnetic study would give a more precise picture but then again interpretation as well as cost would not make it so simple. 

@Rakesh Ganguly- True! Whatever I have proposed is valid for soluble metal complexes. Composite or cluster may have some problems regarding NMR or EPR. Solid state NMR may help!

SQUID can help but your points will be valid for that.

Thanks,

Bye!

Your cluster is oxidizing something, possibly the solvent. How did you detect that it went from Co(III) to Co(II)?

Addition of ascorbic acid may help you to so.This is the way of Fe(iii) reduction to Fe(ii)