We present an analysis searching for dual active galactic nuclei (AGN) among 62 high-redshift (2.5 < z < 3.5) X-ray sources selected from the X-UDS, AEGIS-XD, CDF-S, and COSMOS-Legacy Chandra surveys. We aim to quantify the frequency of dual AGN in the high-redshift Universe, which holds implications for black hole merger timescales and low-frequency gravitational wave detection rates. We analyze each X-ray source using BAYMAX, an analysis tool that calculates the Bayes factor for whether a given archival Chandra AGN is more likely a single or dual point source. We find no strong evidence for dual AGN in any individual source in our sample. We increase our sensitivity to search for dual AGN across the sample by comparing our measured distribution of Bayes factors to that expected from a sample composed entirely of single point sources and find no evidence for dual AGN in the sample distribution. Although our analysis utilizes one of the largest Chandra catalogs of high-z X-ray point sources available to study, the findings remain limited by the modest number of sources observed at the highest spatial resolution with Chandra and the typical count rates of the detected sources. Our nondetection allows us to place an upper limit on the X-ray dual AGN fraction at 2.5 < z < 3.5 of 4.8% at the 95% confidence level. Expanding substantially on these results at X-ray wavelengths will require future surveys spanning larger sky areas and extending to fainter fluxes than has been possible with Chandra. We illustrate the potential of the AXIS mission concept in this regard.