We present the survey design, implementation, and outlook for COSMOS-Web, a 255 hr treasury program conducted by the James Webb Space Telescope in its first cycle of observations. COSMOS-Web is a contiguous 0.54 deg2 NIRCam imaging survey in four filters (F115W, F150W, F277W, and F444W) that will reach 5σ point-source depths ranging ∼27.5–28.2 mag. In parallel, we will obtain 0.19 deg2 of MIRI imaging in one filter (F770W) reaching 5σ point-source depths of ∼25.3–26.0 mag. COSMOS-Web will build on the rich heritage of multiwavelength observations and data products available in the COSMOS field. The design of COSMOS-Web is motivated by three primary science goals: (1) to discover thousands of galaxies in the Epoch of Reionization (6 ≲ z ≲ 11) and map reionization’s spatial distribution, environments, and drivers on scales sufficiently large to mitigate cosmic variance, (2) to identify hundreds of rare quiescent galaxies at z > 4 and place constraints on the formation of the universe’s most-massive galaxies (M⋆ > 1010M⊙), and (3) directly measure the evolution of the stellar-mass-to-halo-mass relation using weak gravitational lensing out to z ∼ 2.5 and measure its variance with galaxies’ star formation histories and morphologies. In addition, we anticipate COSMOS-Web’s legacy value to reach far beyond these scientific goals, touching many other areas of astrophysics, such as the identification of the first direct collapse black hole candidates, ultracool subdwarf stars in the Galactic halo, and possibly the identification of z > 10 pair-instability supernovae. In this paper we provide an overview of the survey’s key measurements, specifications, goals, and prospects for new discovery.
We study the stellar mass assembly of galaxies via the stellar mass function (SMF) and the coevolution with dark matter halos via abundance matching in the largest redshift range to date, 0.2 < z < 12. We used the 0.53 deg2 imaged by JWST from the COSMOS-Web survey, in combination with ancillary imaging in over 30 photometric bands, to select highly complete samples (down to log M⋆/M⊙ = 7.5 − 8.8) in 15 redshift bins. Our results show that the normalization of the SMF monotonically decreases from z = 0.2 to z = 12 with strong mass-dependent evolution. At z > 5, we find increased abundances of massive (log M⋆/M⊙ > 10.5) systems compared to predictions from semi-analytical models and hydrodynamical simulations. These findings challenge traditional galaxy formation models by implying integrated star formation efficiencies (SFEs) of ϵ ⋆ ≡M⋆fb − 1 Mhalo−1 ≳ 0.5. We find a flattening of the SMF at the high-mass end that is better described by a double power law at z > 5.5, after correcting for the Eddington bias. At z ≲ 5.5, it transitions to a Schechter law, which coincides with the emergence of the first massive quiescent galaxies in the Universe, indicating that physical mechanisms that suppress galaxy growth start to take place at z ∼ 5.5 on a global scale. By integrating the SMF, we trace the cosmic stellar mass density and infer the star formation rate density, which at z > 7.5 agrees remarkably with recent JWST UV luminosity function-derived estimates. This agreement solidifies the emerging picture of rapid galaxy formation leading to increased abundances of bright and massive galaxies in the first ∼0.7 Gyr. However, at z ≲ 3.5, we find significant tension (∼0.3 dex) with the cosmic star formation (SF) history from instantaneous SF measures, the causes of which remain poorly understood. We infer the stellar-to-halo mass relation (SHMR) and the SFE from abundance matching out to z = 12, finding a non-monotonic evolution. The SFE has the characteristic strong dependence with mass in the range of 0.02 − 0.2, and mildly decreases at the low-mass end out to z ∼ 3.5. At z ∼ 3.5, there is an upturn and the SFE increases sharply from ∼0.1 to approach a high SFE of 0.8 − 1 by z ∼ 10 for log(Mh/M⊙)≈11.5, albeit with large uncertainties. Finally, we use the SHMR to track the SFE and stellar mass growth throughout the halo history and find that they do not grow at the same rate – from the earliest times up until z ∼ 3.5 the halo growth rate outpaces galaxy assembly, but at z > 3.5 halo growth stagnates and accumulated gas reservoirs keep the SF going and galaxies outpace halos.
We present COSMOS2025, the COSMOS-Web catalog of photometry, morphology, photometric redshifts, and physical parameters for more than 700 000 galaxies in the Cosmic Evolution Survey (COSMOS) field. This catalog is based on our James Webb Space Telescope 255 h COSMOS-Web program, which provides deep near-infrared imaging in four NIRCam (F115W, F150W, F277W, F444W) and one MIRI (F770W) filter over the central ~0.54 deg 2 (~0.2 deg 2 for MIRI) in COSMOS. These data are combined with ground- and space-based data to derive photometric measurements of NIRCam-detected sources using both fixed-aperture photometry (on the space-based bands) and a profile-fitting technique on all 37 bands spanning 0.3 μm to 8 μm. We provide morphology for all sources from complementary techniques including profile fitting and machine-learning classification. We derive photometric redshifts, physical parameters, and non-parametric star formation histories from spectral energy distribution (SED) fitting. The catalog has been extensively validated against previous COSMOS catalogs and other surveys. Photometric redshift accuracy measured using spectroscopically confirmed galaxies out to z ~ 9 reaches σMAD = 0.012 at mF444W < 28 and remains at σMAD ≲ 0.03 as a function of magnitude, color, and galaxy type. This represents a factor of ~2 improvement at 26 AB mag compared to COSMOS2020. The catalog is approximately 80% complete at log( M⋆ /M ⊙ ) ~ 9 at z ~ 10 and at log( M⋆ /M ⊙ ) ~ 7 at z ~ 0.2, representing a gain of 1 dex compared to COSMOS2020. COSMOS2025 represents the definitive COSMOS-Web catalog. It is provided with complete documentation, together with redshift probability distributions, and it is ready for scientific exploitation today.
1 December 2025 in The Astrophysical Journal Supplement Series
We measure galaxy sizes from 2 < z < 10 using COSMOS-Web, the largest-area JWST imaging survey to date, covering ∼0.54 deg 2 . We analyze the rest-frame optical (∼5000 Å) size evolution and its scaling relation with stellar mass ( ) for star-forming and quiescent galaxies. For star-forming galaxies, the slope α remains approximately 0.20 at 2 < z < 8, showing no significant evolution over this redshift range. At higher redshifts, the slopes are −0.13 ± 0.15 and 0.37 ± 0.36 for 8 < z < 9 and 9 < z < 10, respectively. At fixed galaxy mass, the size evolution for star-forming galaxies follows Re ∝ (1 + z ) −1.21±0.05 . For quiescent galaxies, the size–mass relation is steeper with α ∼ 0.5–0.8 at 2 < z < 5, and size evolves as Re ∝ (1 + z ) −0.81±0.26 . We find that the size−mass relation is consistent between UV and optical at z < 8 for star-forming galaxies. However, we observe a decrease in the slope from UV to optical at z > 8, with a tentative negative slope in the optical at 8 < z < 9, suggesting a complex interplay between intrinsic galaxy properties and observational effects such as dust attenuation. We discuss the ratio between galaxies’ half-light radius and underlying halos’ virial radius, Rvir , and find the median value of Re / Rvir = 2.7%. The star formation rate surface density evolves as (0.20 ± 0.08) z + (−0.65 ± 0.51), and the Σ SFR – M* relation remains flat at 2 < z < 10. Lastly, we identify a threshold in stellar mass surface density marking the transition to compact, quenched galaxies from extended, star-forming progenitors. In summary, our findings show that the extensive COSMOS-Web dataset at z > 3 provides new insights into galaxy size and related properties in the rest-frame optical.
Context. Radio used as a tracer of the star formation rate (SFR) presents enormous advantages because it is not affected by dust and radio sources that are located at the subarcsecond level. The interpretation of the low-frequency 1.4 GHz luminosity is hampered by the difficulty of modeling the paths of cosmic rays in the interstellar medium, however, and by their interactions with the magnetic field. Aims. We compare the SFR derived from radio observations and the SFRs derived from spectral energy distribution (SED) modeling. We aim at better understanding the behavior of the SFR radio tracer, with a specific emphasis on the link to star formation histories (SFHs). Methods. The analysis is based on a subsample of 1584 star-forming galaxies extracted from the Cosmic Evolution Survey (COSMOS) with observations of the Very large array project at 3 GHz. We used the SED modeling code investigating galaxy emission, CIGALE, with a nonparametric model for the SFH and fit the data over the wavelength range from the ultraviolet (UV) to the mid-infrared (mid-IR). We interpret the difference between radio and SED-based SFR tracers in the light of recent gradients in the derived SFH. To validate the robustness of the results, we searched for any remaining contribution of active galaxy nuclei and tested the impact of our SFH modeling approach. Results. Approximately 27% our galaxies present a radio SFR (SFRradio) that is at least ten times higher than the instantaneous SFR from SED fitting (SFRSED). This trend primarily affects the galaxies whose SFH activity decreased over the last 300 Myr. Both SFR indicators converge toward a consistent value when the SFHs are averaged over a period longer than 150 Myr to derive SFRSED. Conclusions. Although the radio at a low frequency of 1.4 GHz is a good tracer of the star formation activity of galaxies with a constant or increasing SFH, our results indicate that this is not the case for quenched galaxies. Our analysis suggests that the star formation time sensitivity of the low radio frequency might be longer than 150 Myr. Interestingly, the discrepancy between the SFRradio and SFRSED can be used as diagnostic to select post-starburst galaxies.
We explore the evolving relationship between galaxies and their dark matter halos from z ∼ 0.1 to z ∼ 12 using mass-limited angular clustering measurements in the 0.54 deg2 of the COSMOS-Web survey, the largest contiguous JWST extragalactic survey. This study provides the first measurements of the mass-limited two-point correlation function at z ≥ 10 and a consistent analysis spanning 13.4 Gyr of cosmic history, setting new benchmarks for future simulations and models. Using a halo occupation distribution (HOD) framework, we derived characteristic halo masses and the stellar-to-halo mass ratio (SHMR) across redshifts and stellar mass bins. Our results first indicate that HOD models fit data at z ≥ 2.5 best when incorporating a nonlinear scale-dependent halo bias, boosting clustering at nonlinear scales (r = 10 − 100 kpc). We find that galaxies at z ≥ 10.5 with log(M⋆/M⊙)≥8.85 are predominantly central galaxies in halos with Mh ∼ 1010.5M⊙, achieving a star formation efficiency (SFE) of εSF = M⋆/(fbMh) up to 1 dex higher than at z ≤ 1. The high galaxy bias at z ≥ 8 suggests that these galaxies reside in massive halos with an intrinsic high SFE, challenging stochastic SHMR scenarios. Our SHMR evolves significantly with redshift, starting very high at z ≥ 10.5, decreasing until z ∼ 2 − 3, then increasing again until the present. Current hydrodynamical simulations fail to reproduce both massive high-z galaxies and this evolution, while semi-empirical models linking SFE to halo mass, accretion rates, and redshift align with our findings. We propose that early galaxies (z > 8) experience bursty star formation without significant feedback altering their growth, driving the rapid growth of massive galaxies observed by JWST. Over time, the increasing feedback efficiency and the exponential halo growth end up suppressing star formation. At z ∼ 2 − 3 and later, the halo growth slows down, while star formation continues, supported by gas reservoirs in halos.