Is DirMoE superseded?

DirMoE (Mixture-of-experts routing): current frontier — recent, not yet superseded in the knowledge base. 0 paper(s) critique it, 0 beat it on benchmarks — not ranked as a superseded baseline. Sub-problem: cluster led by DeepSeekMoE. Newer alternatives in the same sub-problem include ProbMoE, Grouter, DECO, SPHERE, AIR-MoE.

Mixture-of-experts routing

DirMoE

DirMoE: Dirichlet-routed Mixture of Experts

On the frontierfirst seen Feb 9, 2026

Current frontier — recent, not yet superseded in the knowledge base

0 papers critique it · 0 beat it on benchmarks

What to use instead

Recent methods in the same sub-problem, not yet superseded in the knowledge base — arXiv benchmark leaders, not vetted production recommendations.

ProbMoEProbMoE: Differentiable Probabilistic Routing for Mixture-of-Experts
Jun 1, 2026
GrouterGrouter: Decoupling Routing from Representation for Accelerated MoE Training
May 24, 2026
DECODECO: Sparse Mixture-of-Experts with Dense-Comparable Performance on End-Side Devices
May 11, 2026
SPHERESPHERE: Mitigating the Loss of Spectral Plasticity in Mixture-of-Experts for Deep Reinforcement Learning
May 6, 2026
AIR-MoEAdaptive Inverted-Index Routing for Granular Mixtures-of-Experts
May 6, 2026
TGR-MoETeacher-Guided Routing for Sparse Vision Mixture-of-Experts
Apr 23, 2026
MalMoEMalMoE: Mixture-of-Experts Enhanced Encrypted Malicious Traffic Detection Under Graph Drift
Feb 10, 2026
DirMoEDirMoE: Dirichlet-routed Mixture of Experts
Feb 9, 2026
OmniMoEOmniMoE: An Efficient MoE by Orchestrating Atomic Experts at Scale
Feb 5, 2026
GRIP (Geometric Routing Invariance Preservation)GRIP: Algorithm-Agnostic Machine Unlearning for Mixture-of-Experts via Geometric Router Constraints
Jan 23, 2026
SR-MoESpectral Manifold Regularization for Stable and Modular Routing in Deep MoE Architectures
Jan 7, 2026