The boron isotopic ratio (δ¹¹B) of marine biogenic carbonates is a proxy for ocean pH through the geological times, provided that δ¹¹B measurements reach precisions better than 0.8‰ (2 SD), corresponding to 0.1 pH-unit (magnitude of the Anthropocene ocean acidification). This level of precision is however challenging to achieve for samples with a small size and low B content used in paleoceanographic reconstructions. We developed a protocol combining (i) the fast-handling and very low blank (9 ± 7 pg, 1 SD, n = 7) microsublimation technique to purify boron from the carbonate matrix and (ii) the direct injection of the samples into the MC-ICP-MS plasma with the new μ-dDIHEN device, allowing for small sample volumes and low uptake rates. We further used Jet sampler and X skimmer cones and implemented two 10¹³ Ω amplifiers to respectively increase the sensitivity and enhance the signal/noise ratio, enabling the precise measurement of small B signals. Our protocol, that consumes only 240 μL for one triplicate sample-standard bracketing (SSB) measurement, was validated through repeated B extractions and measurements of the carbonate reference material MVS-1 at different B concentrations. It yielded 15.93 ± 0.24‰ (2 SD, n = 10) with only 1.2 ng consumed for one triplicate SSB measurement, with individual external repeatabilities (n = 5, 2 SD) of 0.19 ± 0.14‰. We further measured δ¹¹B of modern biogenic carbonates (arthropods, brachiopods, molluscs, red algae and reef carbonate mixes) down to 0.5–0.7 ng of B and successfully compared these δ¹¹B values with those obtained after B extraction of the same samples through ion exchange chromatography.