Accurate prediction of TCR reactivity forms a holy grail in immunology and large language models and computational structure predictions provide a path to achieve this. Importantly, current TCR-pMHC prediction models have been trained and evaluated using historical data of unknown quality.

Here, we develop and utilize a high-throughput synthetic platform for TCR assembly and evaluation to experimentally assess claimed reactivity of VDJdb-deposited TCRs for 8 well-studied viral epitopes, jointly forming approx. 40% of the TCR-pMHC pairs in this database, using a standardized readout of TCR function. Strikingly, this analysis demonstrates that claimed TCR reactivity is only confirmed for approximately 50% of evaluated entries. Intriguingly, the use of TCRbridge to analyze AlphaFold3 confidence metrics for these TCR-pMHC pairs reveals a substantial performance in distinguishing functionally validating and non-validating TCRs even though AlphaFold3 was not trained on this task. These data demonstrate the utility of the validated VDJdb (TCRvdb) database that we generated, and we provide TCRvdb as a resource to the community to support training and evaluation of improved predictive TCR specificity models.