MolMIM
NVIDIA
Molecular generation model using Mutual Information Machine with a Perceiver encoder. Maps molecules into a smooth latent space enabling controlled interpolation and optimization.
Best For
Latent space molecular optimization; lead hopping via smooth interpolation between molecules
License
NVIDIA NIM (API access)
Strengths
- +Smooth latent space for optimization
- +Available as NVIDIA NIM
- +CMA-ES guided generation
- +Fast inference
Limitations
- −Proprietary (API-only via BioNeMo)
- −No local deployment outside NIM
- −SMILES-based (not 3D-aware)
R&D Pipeline Coverage
Related Tools
BioNeMo
NVIDIA
End-to-end AI platform for drug discovery providing GPU-accelerated NIMs (NVIDIA Inference Microservices) spanning protein structure, molecular generation, docking, and property prediction. Includes ESMFold, DiffDock, MolMIM, and 25+ healthcare NIMs.
REINVENT4
AstraZeneca Molecular AI
RL + transformer platform for de novo small molecule design. Supports scaffold decoration, R-group replacement, linker design, and multi-parameter optimization.
SAFE-GPT
Valence Labs
GPT-style model trained on SAFE (fragment-based) molecular representation. Enables fragment-constrained design including scaffold decoration and linker generation.
More in Generative Design
RFdiffusion
Baker Lab / IPD (University of Washington)
Diffusion-based generative model for de novo protein backbone design. Generates novel protein structures conditioned on binding targets, symmetry, or functional sites.
RFdiffusion2
Baker Lab / IPD (University of Washington)
Successor to RFdiffusion using flow matching. Designs enzymes directly from active site geometry (theozyme) specifications.
ProteinMPNN
Baker Lab / IPD (University of Washington)
Inverse folding model: generates amino acid sequences predicted to fold into a target 3D backbone structure. Standard component of all modern protein design pipelines.
Stay updated on MolMIM
Weekly newsletter covering AI tool releases, benchmarks, and what practitioners actually use.