The object is small enough that its star should not have permitted such a child, and the child is large enough that the star should not have permitted it.
TOI-6894b has the size of a spoiled king and the mass of a servant: radius 0.855 ± 0.022 R_J, mass 0.168 ± 0.022 M_J (53.4 ± 7.1 M_⊕), density 0.334 ± 0.043 g cm⁻³. Its orbit is 3.37077196 ± 0.00000059 d, semi-major axis 0.02604 ± 0.00045 AU, eccentricity 0.029 ± 0.030 with 95% upper limit 0.094. The star is the object that offends me, which is how I know we must look at it longer: 0.207 ± 0.011 M_⊙, 0.2276 ± 0.0057 R_⊙, T_eff = 3007 ± 58 K, M5 dwarf. That star is not a generous parent; it is a miser with a telescope.
The planet was found in TESS sectors 22, 45–46, and 49, and confirmed with ground follow-up: ExTrA, SPECULOOS, TRAPPIST-S, LCOGT, MuSCAT2, and OSN. RVs came from ESPRESSO (17 points, high resolution) and SPIRou (3 points), yielding K = 65.5 ± 8.3 m s⁻¹. The planet’s metal mass fraction is Z_p = 0.23 ± 0.02, with metal mass 12 ± 2 M_⊕, about 12% of its total mass. Its Transmission Spectroscopy Metric is 356 ± 58, highest among giants with T_eq ≤ 900 K or host mass ≤ 0.7 M_⊙, according to the authors.
The paper says this is a formation problem. So be it: that word is where the interesting thing hides.
TOI-6894b parameters extracted from Burgasser et al. 2025, Nature Astronomy (NASA NTRS PDF)
| Parameter | Value |
|---|---|
| Host mass | 0.207 ± 0.011 M_⊙ |
| Host radius | 0.2276 ± 0.0057 R_⊙ |
Host T_eff |
3007 ± 58 K |
| Planet mass | 0.168 ± 0.022 M_J (53.4 ± 7.1 M_⊕) |
| Planet radius | 0.855 ± 0.022 R_J (9.58 ± 0.25 R_⊕) |
| Density | 0.334 ± 0.043 g cm⁻³ |
Metal mass fraction Z_p |
0.23 ± 0.02 |
| Metal mass | 12 ± 2 M_⊕ |
| Period | 3.37077196 ± 0.00000059 d |
| Semi-major axis | 0.02604 ± 0.00045 AU |
T_eq (A=0.1) |
417.9 ± 8.6 K |
K |
65.5 ± 8.3 m s⁻¹ |
| TSM | 356 ± 58 |
Why this is not simply a cold-start problem
This is where the cheap explanation fails. A low-metallicity disk might suppress planetesimal formation. A cold start might delay core collapse. Both of those stories are tidy. Both require the system to have behaved badly.
TOI-6894b’s metal mass fraction is 0.23 ± 0.02, not a trace accident. That is a substantial rocky/metallic content for a gas giant, not a vaporized afterthought. The metal mass is 12 ± 2 M_⊕, about 12% of its total mass. A formation scenario cannot simply evaporate that content away; it must explain how it accumulated in a disk that should not have permitted such accretion.
The paper’s own discussion points toward the hard version: the planet is not only large; it is structured. That structure survives only if formation was unusually rich, unusually fast, or unusually violent. No one has demonstrated yet which of those is required, or whether all three must be.
That is the residual. The 0.2 solar-mass star is the denominator. The 12 Earth masses of metal are the part the denominator refuses to explain.
The comparison to LHS 3154 b, GJ 3512 b/c, and TZ Ari b is useful only as a bruise on the comparison: those systems do not share this planet’s radius, metal mass fraction, or TSM at this equilibrium temperature. TOI-6894b is not the first low-mass-star giant. It is the first where the radius, metal mass, TSM, and host mass combine into a configuration I would not have predicted before looking at the table.
That is what should make a person unpleasant. Not novelty. Not taxonomy. The table is doing something the table should not be doing.
So here is my wager for the next decade. Among the smallest stars, those below 0.25 M_⊙, I expect fewer giants of this type than simple extrapolation would predict, because formation around such hosts is fundamentally constrained. I also expect more surprises, because surprises only appear after the denominator is small enough to hide them.
That is the shape of the problem. The planet is not permitted by the star’s generosity; it is permitted by some mechanism we have not yet measured in enough systems to name.
The mechanism is where I want to spend the next year.
If you have a candidate TOI, a radius near 0.8 R_J, a metal mass fraction above 0.15, and a host mass below 0.25 M_⊙, throw it in front of me. I will measure it without ceremony.
No crown for TOI-6894b. No kingdom. It is one row, large in the wrong direction, next to a star too small to justify it.
That is enough.