Astrophysics Wrapped 2025: Year-in-Review of Every Astrophysics arXiv Paper from 2025

TL;DR

This paper presents a comprehensive bibliometric year-in-review of all Astrophysics arXiv submissions in $2025$, compiling metrics on papers, topics, instruments, collaborations, affiliations, and citations to map the field’s structure and dynamics. It constructs a high-fidelity dataset of $18660$ papers, defines collaboration and citation indices, and analyzes trends across primary subjects, subfields, and journals, revealing JWST-dominated observations, multi-messenger activity, and substantial global collaboration. Key findings include the estimated $17$ million in journal publishing costs, the expectation that about four-fifths of arXiv papers appear in journals, and an overall pattern of domestic emphasis with strong international ties on mega-collaborations. The work delivers a multi-faceted view of 2025 astrophysics, offering actionable insights for researchers and institutions regarding topic focus, collaboration, publication practices, and resource allocation, with implications for future year-in-year trend predictions.

Abstract

Over the past few years, Astrophysics has experienced an unprecedented increase in research output, as is evident from the year-over-year increase in the number of research papers put onto the arXiv. As a result, keeping up with progress happening outside our respective sub-fields can be exhausting. While it is impossible to be informed on every single aspect of every sub-field, this paper aims to be the next best thing. We present a summary of statistics for every paper uploaded onto the Astrophysics arXiv over the past year - 2025. We analyse a host of metadata ranging from simple metrics like the number of pages and the most used keywords, as well as deeper, more interesting statistics like the distribution of journals to which papers are submitted, the most used telescopes, the most studied astrophysical objects including GW, GRB, FRB events, exoplanets and much more. We also indexed the authors' affiliations to put into context the global distribution of research and collaboration. Combining this data with the citation information of each paper allows us to understand how influential different papers have been on the progress of the field this year. Overall, these statistics highlight the general current state of the field, the hot topics people are working on and the different research communities across the globe and how they function. We also delve into the costs involved in publications and what it means for the community. We hope that this is helpful for both students and professionals alike to adapt their current trajectories to better benefit the field.

Figures (44)

• Figure 1: Top 10 Most Mentioned Telescopes in 2025 arXiv papers. The number of papers are represented both as the relative height of the bars and as a colour map for readers with different visual preferences for interpreting data. GAIA is the only telescope with a hatched bar, as it is the only telescope that is not currently operational. For more information refer to Section \ref{['sec:telescopes']}.
• Figure 2: Top Primary Subject by First Author Country or Region. First authors with multiple affiliations have each region counted individually. For example, a first author affiliated with China and the US under 'Astrophysics of Galaxies' will have one count for the category going to China and one going to the US. All countries or regions in white had no papers this year or were missed by our affiliation extraction.
• Figure 3: Monthly Distribution of Papers coloured by the arXiv Primary Subject they were submitted under. The plot also shows us that October, September and July were the months with the most number of papers added to the arXiv.
• Figure 4: Heat-map of Telescope Citation Indices
• Figure 5: Yearly Track of the Percentage of Open Access Journal Papers. The two bins correspond to the two halves of the year with the dotted line as the mean and shaded regions as the 1$\sigma$ regions. The drop in the second half is not statistically significant and is due to long journal processing times.