Sachdev–Ye–Kitaev (SYK) Model

• The SYK model is a zero-dimensional many-body system of Majorana fermions with random, all-to-all q-body interactions, leading to a non-Fermi liquid behavior.
• It features a large-N bilocal formulation that reveals an emergent infrared conformal symmetry and a Schwarzian soft mode driving maximal quantum chaos.
• Its implications extend to quantum gravity and holography, linking near-AdS2 dynamics, thermodynamics with residual entropy, and experimental quantum simulation proposals.

Searching arXiv for recent and foundational SYK references to support the encyclopedia entry. arXiv search query: "Sachdev-Ye-Kitaev model review conformal Schwarzian chaos large N" The Sachdev–Ye–Kitaev (SYK) model is a zero-dimensional quantum many-body model of fermions with random, all-to-all interactions. In its standard form it is written for a large number of Majorana fermions with even-$q$ interactions, and in the large-$N$ limit it is solvable in terms of bilocal collective fields. Its importance rests on the coexistence of several otherwise unusual features in a single framework: strong interaction without quasiparticles, emergent infrared conformal structure, a Schwarzian soft mode, maximal quantum chaos diagnosed by out-of-time-ordered correlators (OTOCs), finite residual entropy density, and a close relation to near-$AdS_2$ black-hole dynamics (Maldacena et al., 2016, Sachdev, 2024, Jha, 9 Jul 2025).

1. Definition and canonical formulations

The canonical Majorana SYK model consists of $N$ Majorana fermions $\psi_i$ in $0+1$ dimensions with random all-to-all $q$-body interactions, most often specialized to $q=4$. A standard definition is

$$H = i^{q/2}\sum_{1\le i_1<i_2<\cdots<i_q\le N} j_{i_1 i_2\cdots i_q}\,\psi_{i_1}\psi_{i_2}\cdots \psi_{i_q},$$

with

$$\{\psi_i,\psi_j\}=\delta_{ij},\qquad \langle j_{i_1\cdots i_q}\rangle =0,\qquad \langle j_{i_1\cdots i_q}^2\rangle = \frac{J^2 (q-1)!}{N^{q-1}}.$$

The $N$0-dependence of the variance is essential because it yields a nontrivial large-$N$1 limit with extensive interaction energy (Maldacena et al., 2016).

A closely related charge-conserving formulation uses complex fermions. In the notation of one survey,

$N$2

with

$N$3

and conserved charge density

$N$4

This version retains the defining SYK ingredients—randomness, all-to-all quartic interactions, and large-$N$5 solvability—while making the $N$6 sector explicit (Sachdev, 2024).

The model is “zero-dimensional” in the precise sense that it has no spatial locality and no momentum-space structure. This absence of locality is not incidental: it is part of why the model evades conventional quasiparticle descriptions and instead organizes its infrared dynamics through bilocal collective fields rather than single-particle modes (Jha, 9 Jul 2025).

2. Large-$N$7 bilocal formulation and infrared conformal regime

After disorder averaging, the large-$N$8 description is expressed in terms of bilocal fields

$N$9

and a conjugate self-energy $AdS_2$0. In the Majorana case, a standard effective action is

$AdS_2$1

and the saddle-point equations are the Schwinger–Dyson equations

$AdS_2$2

for the $AdS_2$3 model, or more generally

$AdS_2$4

These equations resum the melonic large-$AdS_2$5 expansion (Rosenhaus, 2018, Jha, 9 Jul 2025).

In the infrared, the kinetic term becomes subleading and the equations acquire an emergent reparameterization symmetry. The conformal ansatz

$AdS_2$6

solves the infrared equations with

$AdS_2$7

so for the canonical $AdS_2$8 model the fermion scaling dimension is $AdS_2$9. At finite temperature the conformal solution is obtained by the reparameterization $N$0, which gives the standard low-temperature SYK form (Polchinski et al., 2016, Jha, 9 Jul 2025).

The emergent symmetry is not exact. In the infrared the theory is approximately invariant under $N$1, with

$N$2

but the conformal saddle preserves only the $N$3 subgroup. The resulting pseudo-Goldstone mode is governed by the Schwarzian action

$N$4

This Schwarzian sector controls the very low-energy dynamics and is the basic bridge between SYK and nearly-$N$5 gravity (Maldacena et al., 2016).

3. Four-point structure, bilocal spectrum, and maximal chaos

The leading connected four-point function is of order $N$6 and is obtained by summing ladder diagrams. In one standard representation the ladder kernel is

$N$7

so the four-point problem reduces to diagonalizing $N$8 and summing $N$9 against the zero-rung ladder (Maldacena et al., 2016).

In the large-$\psi_i$0 spectral analysis of the two-particle sector, the kernel admits both a continuous and a discrete tower. One convenient parametrization uses eigenfunctions $\psi_i$1, with continuous states

$\psi_i$2

and discrete states

$\psi_i$3

The connected four-point function can then be expressed as a spectral sum over these modes. This decomposition gives a precise operator-theoretic content to the statement that SYK is solvable at large $\psi_i$4: the bilocal field has an explicit excitation spectrum rather than a purely formal collective description (Polchinski et al., 2016).

In the conformal language, the four-point function is decomposed into conformal blocks, and the dimensions of bilinear singlet operators are determined by the equation $\psi_i$5, where $\psi_i$6 is the conformal kernel eigenvalue. The $\psi_i$7 sector is special. In the strict conformal limit it sits at the edge of divergence, and the small explicit breaking of reparameterization symmetry lifts this mode and produces an enhanced contribution to the four-point function. That enhancement is the origin of the dominant chaotic behavior (Maldacena et al., 2016).

The chaos diagnostic is the OTOC. The general bound

$\psi_i$8

is saturated by SYK at low temperatures,

$\psi_i$9

so the model is maximally chaotic in the Maldacena–Shenker–Stanford sense. This is not an isolated kinematic statement: it is directly tied to the Schwarzian $0+1$0 mode, to the emergent reparameterization structure, and to the ladder-kernel organization of the four-point function (Rosenhaus, 2018, Jha, 9 Jul 2025).

4. Thermodynamics, spectral statistics, and holographic interpretation

A central thermodynamic feature is the residual entropy. In the large-$0+1$1 limit the low-temperature entropy may be written as

$0+1$2

so that

$0+1$3

as $0+1$4 after taking $0+1$5. The important distinction is that this extensive zero-temperature entropy does not require an exponentially large exact ground-state degeneracy. Rather, the many-body level spacing near the ground state is exponentially small in $0+1$6, and the entropy comes from a dense pileup of low-energy states above a unique finite-$0+1$7 ground state (Sachdev, 2024).

The same survey emphasizes that the infrared spectral functions are not those of bosons or fermions in the quasiparticle sense but are instead “Planckian,” meaning they are universal functions of energy/temperature. Correspondingly, the relaxation timescale obeys

$0+1$8

and the infrared state is a quantum critical liquid rather than a glass. This distinction is important because SYK is often grouped with random infinite-range systems, yet its equilibrium infrared dynamics are conformal rather than glassy (Sachdev, 2024).

The holographic connection is organized around the near-horizon $0+1$9 region of charged or rotating black holes. The Euclidean metric

$q$0

has $q$1 isometry, matching the infrared symmetry of the SYK saddle. The low-energy effective theory is described by time reparameterizations $q$2 and, in the charged case, a $q$3 phase mode $q$4, with the Schwarzian action appearing on both the SYK and Jackiw–Teitelboim sides. In this precise low-energy sense, SYK supplies an effective theory for near-extremal black-hole dynamics rather than merely an analogy (Sachdev, 2024).

Beyond formal large-$q$5 physics, part of the equilibrium thermodynamics is now rigorously controlled. For even $q$6, fixed sufficiently small $q$7, and $q$8, the annealed free energy density has a deterministic limit, and prior work cited there implies that the quenched and annealed free energies coincide in that regime. The proof uses sparse random hypergraph component analysis and a cavity-method-style recursion rather than the standard replica/path-integral derivation (Gamarnik et al., 4 May 2026).

Spectral statistics provide a complementary signature of chaos. In the original complex SYK parameter regime examined in one study, the level statistics are GUE-like, consistent with quantum chaos, whereas the unitary complex SYK variant studied in another work agrees with GOE expectations. The same pruning study shows that short-time chaotic behavior can survive a drastic reduction of couplings: keeping only $q$9 quartic terms with $q=4$0, rather than the full $q=4$1 set, still preserves GOE ratio statistics, an RMT-like SVD scree-plot regime, and GOE-like level number variance over a substantial short-time window (Iyoda et al., 2018, Berkovits, 2024).

5. Variants, deformations, and generalized SYK constructions

The label “SYK-like” covers models with very different physics, and several constructions make explicit what is and is not universal. A notable example is the Wishart SYK model,

$q=4$2

Although it is still a disordered four-fermion model, its fermionic version is integrable: after a canonical reduction of the antisymmetric coupling matrix it becomes a Richardson–Gaudin model with mutually commuting conserved charges. Its level statistics are Poisson rather than Wigner–Dyson, its ground state is massively degenerate, and its late-time OTOC displays large temporal fluctuations associated with a small effective dimension. This is a direct counterexample to the idea that any all-to-all random quartic model automatically inherits SYK chaos (Iyoda et al., 2018).

Another deformation retains infinite-range disorder but changes the disorder tensor structure. The two-index Sachdev–Ye–Kitaev model rewrites the original Sachdev–Ye spin glass in a Majorana representation with only $q=4$3 disorder rather than $q=4$4. At $q=4$5, its finite-$q=4$6 quantum spin liquid remains stable, the $q=4$7 corrections are exactly marginal, and the quantum spin glass instability occurs only at the exponentially small scale

$q=4$8

The proposal is motivated in part by the possibility that a two-index disorder structure could fit a bulk string description better than the usual four-index ensemble (Ye, 2018).

SYK ideas also extend into topological band structures. In the topological SYK model, a two-band Chern-insulator hopping term is combined with onsite SYK interactions. The interaction renormalizes the effective mass,

$q=4$9

and can drive a transition from a trivial insulator to a topological phase with quantized Hall conductance. At the critical point a single gapless Dirac fermion emerges. The same work finds that stronger interaction enlarges the topological regime at $$H = i^{q/2}\sum_{1\le i_1<i_2<\cdots<i_q\le N} j_{i_1 i_2\cdots i_q}\,\psi_{i_1}\psi_{i_2}\cdots \psi_{i_q},$$0 but makes the topological phase less stable against temperature (Zhang et al., 2018).

Open-system generalizations replace unitary evolution by a Lindbladian. For SYK coupled to Markovian reservoirs, with either linear jumps $$H = i^{q/2}\sum_{1\le i_1<i_2<\cdots<i_q\le N} j_{i_1 i_2\cdots i_q}\,\psi_{i_1}\psi_{i_2}\cdots \psi_{i_q},$$1 or random quadratic jumps $$H = i^{q/2}\sum_{1\le i_1<i_2<\cdots<i_q\le N} j_{i_1 i_2\cdots i_q}\,\psi_{i_1}\psi_{i_2}\cdots \psi_{i_q},$$2, the large-$$H = i^{q/2}\sum_{1\le i_1<i_2<\cdots<i_q\le N} j_{i_1 i_2\cdots i_q}\,\psi_{i_1}\psi_{i_2}\cdots \psi_{i_q},$$3 stationary dynamics remain analytically tractable. In the linear-jump case the dominant decay rate approaches

$$H = i^{q/2}\sum_{1\le i_1<i_2<\cdots<i_q\le N} j_{i_1 i_2\cdots i_q}\,\psi_{i_1}\psi_{i_2}\cdots \psi_{i_q},$$4

for strong dissipation, while random quadratic jumps yield an underdamped-to-overdamped crossover and Lorentzian spectral functions at large dissipation (Kulkarni et al., 2021).

For applications to correlated materials, a spatial extension is required. The survey on quantum spin glasses and SYK emphasizes that the two-dimensional Yukawa-SYK model is the appropriate extension for strange-metal phenomenology in systems with spatial inhomogeneity, where it yields a universal description of linear-in-$$H = i^{q/2}\sum_{1\le i_1<i_2<\cdots<i_q\le N} j_{i_1 i_2\cdots i_q}\,\psi_{i_1}\psi_{i_2}\cdots \psi_{i_q},$$5 resistivity, anomalous optical conductivity, and marginal-Fermi-liquid-like spectra over an intermediate-energy regime (Sachdev, 2024).

6. Experimental and quantum-simulation realizations

A longstanding challenge is that the ideal SYK Hamiltonian requires dense random all-to-all quartic couplings with negligible bilinears. One early solid-state proposal approximates this structure with a disordered quantum dot strongly coupled to Majorana zero modes at the ends of topological superconducting wires. The dot randomizes the Majorana wavefunctions and mediates effective quartic couplings

$$H = i^{q/2}\sum_{1\le i_1<i_2<\cdots<i_q\le N} j_{i_1 i_2\cdots i_q}\,\psi_{i_1}\psi_{i_2}\cdots \psi_{i_q},$$6

while an approximate time-reversal symmetry suppresses unwanted bilinears $$H = i^{q/2}\sum_{1\le i_1<i_2<\cdots<i_q\le N} j_{i_1 i_2\cdots i_q}\,\psi_{i_1}\psi_{i_2}\cdots \psi_{i_q},$$7. Random-matrix arguments and numerics show that the effective $$H = i^{q/2}\sum_{1\le i_1<i_2<\cdots<i_q\le N} j_{i_1 i_2\cdots i_q}\,\psi_{i_1}\psi_{i_2}\cdots \psi_{i_q},$$8 are approximately Gaussian, with corrections controlled by dot size and symmetry breaking (Chew et al., 2017).

Digital quantum simulation has already accessed small SYK instances. A four-qubit nuclear magnetic resonance experiment encoded $$H = i^{q/2}\sum_{1\le i_1<i_2<\cdots<i_q\le N} j_{i_1 i_2\cdots i_q}\,\psi_{i_1}\psi_{i_2}\cdots \psi_{i_q},$$9 Majorana fermions, implemented the generalized SYK Hamiltonian via Jordan–Wigner mapping and Trotter–Suzuki decomposition, and measured the bosonic pair correlator

$$\{\psi_i,\psi_j\}=\delta_{ij},\qquad \langle j_{i_1\cdots i_q}\rangle =0,\qquad \langle j_{i_1\cdots i_q}^2\rangle = \frac{J^2 (q-1)!}{N^{q-1}}.$$0

It observed the fermion-pairing instability and the chaotic-to-nonchaotic transition predicted for the deformed model, with the pure $$\{\psi_i,\psi_j\}=\delta_{ij},\qquad \langle j_{i_1\cdots i_q}\rangle =0,\qquad \langle j_{i_1\cdots i_q}^2\rangle = \frac{J^2 (q-1)!}{N^{q-1}}.$$1 case showing the fastest decay consistent with a maximally chaotic non-Fermi liquid (Luo et al., 2017).

A distinct route uses purely spin degrees of freedom. In a proposal based on a Kitaev spin chain, localized $$\{\psi_i,\psi_j\}=\delta_{ij},\qquad \langle j_{i_1\cdots i_q}\rangle =0,\qquad \langle j_{i_1\cdots i_q}^2\rangle = \frac{J^2 (q-1)!}{N^{q-1}}.$$2 Majorana zero modes appear because the unperturbed Majorana Hamiltonian contains no $$\{\psi_i,\psi_j\}=\delta_{ij},\qquad \langle j_{i_1\cdots i_q}\rangle =0,\qquad \langle j_{i_1\cdots i_q}^2\rangle = \frac{J^2 (q-1)!}{N^{q-1}}.$$3 terms. A structured perturbation has vanishing first-order contribution within the ground-state manifold, while second-order perturbation theory generates quartic terms

$$\{\psi_i,\psi_j\}=\delta_{ij},\qquad \langle j_{i_1\cdots i_q}\rangle =0,\qquad \langle j_{i_1\cdots i_q}^2\rangle = \frac{J^2 (q-1)!}{N^{q-1}}.$$4

in the low-energy sector, precisely the operator structure required for a $$\{\psi_i,\psi_j\}=\delta_{ij},\qquad \langle j_{i_1\cdots i_q}\rangle =0,\qquad \langle j_{i_1\cdots i_q}^2\rangle = \frac{J^2 (q-1)!}{N^{q-1}}.$$5 SYK Hamiltonian. The same work constructs spin-string correlators that approximately reproduce Majorana OTOCs, thereby giving a bosonic route to SYK-like scrambling diagnostics (Zuo et al., 2024).

Cavity-QED platforms offer another family of implementations. A multimode optical-cavity proposal with fermionic atoms and a spatially disordered AC-Stark shift derives an effective quartic interaction tensor

$$\{\psi_i,\psi_j\}=\delta_{ij},\qquad \langle j_{i_1\cdots i_q}\rangle =0,\qquad \langle j_{i_1\cdots i_q}^2\rangle = \frac{J^2 (q-1)!}{N^{q-1}}.$$6

and finds random-matrix-like spectral properties, dip-ramp-plateau spectral form factors, and fast scrambling even though the couplings are often more Cauchy-like than Gaussian (Uhrich et al., 2023). A related single-mode cavity proposal replaces exact dense disorder by rapid cycling through reduced-rank random layers,

$$\{\psi_i,\psi_j\}=\delta_{ij},\qquad \langle j_{i_1\cdots i_q}\rangle =0,\qquad \langle j_{i_1\cdots i_q}^2\rangle = \frac{J^2 (q-1)!}{N^{q-1}}.$$7

combined with Trotterization. The resulting disorder distribution approaches the dense SYK ensemble, with a Kullback–Leibler divergence

$$\{\psi_i,\psi_j\}=\delta_{ij},\qquad \langle j_{i_1\cdots i_q}\rangle =0,\qquad \langle j_{i_1\cdots i_q}^2\rangle = \frac{J^2 (q-1)!}{N^{q-1}}.$$8

for $$\{\psi_i,\psi_j\}=\delta_{ij},\qquad \langle j_{i_1\cdots i_q}\rangle =0,\qquad \langle j_{i_1\cdots i_q}^2\rangle = \frac{J^2 (q-1)!}{N^{q-1}}.$$9 independent layers (Baumgartner et al., 2024).

Floquet engineering provides a route from local Hubbard physics to SYK-like quartic interactions. In the kinetically driven Bose–Hubbard proposal,

$N$00

so single-particle hopping averages away in the high-frequency limit and the effective Hamiltonian becomes

$N$01

Direct comparison of the spectral form factor and OTOCs with the bosonic SYK model shows that the driven system reproduces SYK diagnostics despite non-Gaussian and not perfectly all-to-all couplings (Creffield et al., 2 Dec 2025).

Near-term hardware realizations also benefit from sparsification. A trapped-ion experiment simulated a sparse SYK model with $N$02 Majorana fermions using the randomized TETRIS algorithm. The Hamiltonian kept only a fraction $N$03 of the quartic terms,

$N$04

and the experiment measured the Loschmidt amplitude long enough to observe its decay, combining randomized evolution with hardware-tailored error mitigation (Granet et al., 10 Jul 2025).

Taken together, these implementations show that present realizations rarely produce the textbook ensemble of fully independent Gaussian $N$05. Instead they generate structured, sparse, low-rank, or perturbatively emergent quartic couplings. A plausible implication is that the experimentally relevant notion of “SYK physics” is often operational—maximal or near-maximal scrambling, nonlocal quartic interactions, random-matrix spectral rigidity, and Schwarzian-like low-energy structure—rather than exact microscopic identity with the ideal disorder ensemble.