arXiv2026
We present a comprehensive multiwavelength analysis of GRB 240825A, a bright gamma-ray burst (GRB) detected by Fermi and Swift, with a prompt duration ($T_{\rm 90}$ $\sim$ 4 sec in 50-300 keV in GBM) near the boundary separating short and long GRBs, prompting a detailed investigation into its classification and progenitor. We use classical prompt metrics (duration, minimum variability timescale (MVT), lag, and spectral hardness) and modern classification techniques (machine-learning (ML) based t-SNE, support vector machine, energy-hardness-duration, and $\varepsilon \equiv E_{γ,\mathrm{iso},52} / E_{p,z,2}^{5/3}$) and find most properties (prompt energetics, placement on the Amati relation, and spectral lag) of GRB 240825A consistent with a collapsar origin. However, extensive late-time optical and NIR follow-up with the 10.4m GTC and 8.4m binocular LBT telescopes reveals no bright supernova (like SN 1998bw) is detected down to stringent limits (e.g., $m_r > 25.0$ mag at 17.59 days), despite a redshift of $z = 0.659$ measured from GTC spectroscopy. Host galaxy SED modeling with Prospector indicates a massive, and star-forming galaxy-typical of collapsar GRB hosts, though with a large offset. We compare these findings with hybrid events like GRB 211211A, GRB 230307A, GRB 200826A, including SNe-GRBs, and conclude that GRB 240825A most likely originated from a massive star collapse, possibly with the associated SN heavily obscured or intrinsically faint. This study emphasizes the need for multiwavelength follow-up and a multi-layered classification to determine GRB progenitors.