Tree rings provide long-term records of tree growth and climate changes, which makes them ideal benchmarks for forest modeling. Tree-ring information has greatly improved the reliability of 3-PG, which is one of the most commonly used process-based forest growth models. Here, we strengthen 3-PG's ability to simulate tree-ring width and stable carbon isotopes (δ¹³C) by enhancing its descriptions of tree physiology. The major upgrade was adding a carbon storage pool for tree-ring formation using stored carbohydrates. We also incorporated previous modifications (replacing the age modifier with a height modifier) of 3-PG and tested their efficacy in improving tree-ring simulations. We ran the model based on two grand fir (Abies grandis) stands. The updated model greatly improved the simulations for both tree-ring widths and δ¹³C. The results represent one of the best tree-ring δ¹³C simulations, which accurately captured the amplitude in annual variations of δ¹³C. The correlations (R²) between simulations and observations reached 0.50 and 0.73 ​at two stands, respectively. The new model also greatly improved the simulations of raw tree-ring widths and detrended ring-width index (RWI). Because of better descriptions of tree physiology and more accurate simulations of tree rings than the previous model version, the updated 3-PG should provide more reliable simulations than previous 3-PG versions when tree-ring information is used as a benchmark in future studies.