#ESMFold
localfold.org updated to include:
- bioemu2 @franknoe.bsky.social
- mol1 @MolenceBio
- fastplm (confidence modules for esmfold fast models) @lhallee.bsky.social
- af3x crosslinks @jankosinski.bsky.social
- faster diffusion (exponential update and geometric noise schedule) Vinay Mathew
November 2, 2026 at 4:17 AM
Tool Duel: AlphaFold 2 vs ESMFold. AF2 searches for related sequences and reads the alignment with a 48-block Evoformer; ESMFold swaps the search for the ESM-2 language model. Both end in AF2's structure module design. By DeepMind & Meta AI; details at toolome.bio #AlphaFold #CompBio
October 26, 2026 at 3:22 PM
Protein is the largest category on Toolome: 54 of 293 tools. MIT 20, Apache-2.0 12, BSD-3-Clause 9, other 13; 2024 had the most first releases (11). Featured: ProteinMPNN, ESM-2 / ESMFold, RFdiffusion, BindCraft, Boltz. toolome.bio
October 18, 2026 at 3:23 PM
Una precisión: la IA generativa sí está revolucionando el plegado de proteínas, pero no todo son LLMs.
AlphaFold 3 genera la estructura por difusión pero no es un LLMs.
ESMFold sí usa un LLM, pero no pliega por sí solo: extrae patrones de la secuencia que luego se convierten en la estructura 3D
October 10, 2026 at 8:53 AM
Drop any AlphaFold 2/3, ColabFold, Boltz-2, Chai-1, OpenFold3, Protenix or ESMFold output (zipped is fine):
- iLIS next to ipSAE, pDockQ2, actifpTM…
- PAE and contact maps, chord diagrams
- Mol* 3D, ChimeraX/PyMOL scripts
- big complexes, reported PPIs (MIST + BioGRID)
September 25, 2026 at 2:44 PM
I set up a website to make it easy to run common structural Bio/ML tools like RFDiffusion, structure prediction, and sequence prediction tools (ESMFold, ProteinMPNN, Chai-1 etc). athanortools.com

I'm using this for a specific protein design project and learning journey.
Athanor Bio Tools
Athanor Bio Tools provides small web interfaces and JSON APIs for computational biology software.
athanortools.com
September 19, 2026 at 4:56 PM
September 10, 2026 at 6:25 PM
Prot-LAMBDA: a protein language model taught to reason in 3D, reportedly matching one 5x its size. Efficiency over scale. Read the breakdown. #ProteinAI #ESMFold #CompBio

artificialscience.org/2026/08/prot...
Protein Language Models Just Learned to Think in 3D — Artificial Science
A new bioRxiv preprint, Prot-LAMBDA, teaches protein language models explicit 3D geometry and reportedly matches a model five times its size. Why efficiency, not scale, is the real story.
artificialscience.org
August 28, 2026 at 5:10 AM
We found earlier that in randomized and de novo emerged proteins predicted disordered regions showed high pLDDT, while predicted beta-sheets actually showed low pLDDT. Using Alphafold2, fldpnn, Iupred3, ESMfold 2/6

onlinelibrary.wiley.com/doi/full/10....
Random, de novo, and conserved proteins: How structure and disorder predictors perform differently
Understanding the emergence and structural characteristics of de novo and random proteins is crucial for unraveling protein evolution and designing novel enzymes. However, experimental determination ...
onlinelibrary.wiley.com
August 26, 2026 at 2:07 PM
I found out that Neurosnap offers ESMFold via API neurosnap.ai/service/ESMF...

I may test how many calls are possible with the free academic plan to see if it is worthwhile to update my repo.
🧬 Use ESMFold Online | Neurosnap
Bulk protein structure prediction model that only requires a single amino acid sequence as input. Much faster than AlphaFold2 since no MSAs are required (but slightly less accurate too).
neurosnap.ai
October 9, 2025 at 2:25 AM
Let’s dive into the world of #AI and #EnzymeTechnology! 🤖
The new AI.zymes platform, developed by @ethzurich.bsky.social and @bristoluni.bsky.social uses advanced protein engineering algorithms (like Rosetta, ESMFold and ProteinMPNN) to optimize enzyme properties for various industries and research
AI.zymes – A modular platform for evolutionary enzyme design
The ability to create new-to-nature enzymes would substantially advance bioengineering, medicine, and the chemical industry. Despite recent breakthroughs in protein design and structure prediction,…
www.biorxiv.org
April 23, 2025 at 10:02 AM
Thanks for asking!
If I recall correctly, it’s ESMfold protein structures.
January 28, 2026 at 9:59 PM
The team first reconstructed a comprehensive phylogeny of 4,854 insects spanning all 28 orders, then used it to select 824 representative species for structure prediction with ESMFold.
January 28, 2026 at 6:57 PM
Genome hasn’t been released yet I believe but I can send you the sequence. We’ve run it in AF2,AF3,boltz2 and ESMfold. Some variability but not wildly different. Electrostatics are wonder to behold. Did find another gene in a Loki with just one of the ‘badge’ domains- looks like a gene duplication.
July 24, 2025 at 10:08 PM
We are folding a genome this weekend with Boltz1 which has been previously folded with ESMfold and also some with AF3 and AF2. I’ll let you know how it goes.
December 15, 2024 at 2:46 AM
And how about solving your structure using Rosettafold? Or ESMfold? How do they prove you used alphafold?
September 25, 2023 at 2:36 PM
For those of you wondering how ESMFold2 from the folks at @biohub.org performs on peptide:MHC complexes, their lovely API is currently running a test set.

As I said before with ESMFold. Make APIs, it makes things easy for people to build on.
May 28, 2026 at 4:05 PM
Boltz-1x
Chai-1
Protenix
AF2
AF3
Unifold
CombFold
ESMFold
D-I-TASSER
RosettaFold All-Atom
IgDesign
RFantibody
BindCraft
RoseTTAFold2NA
AFCluster
AntiFold
HighFold
BioEmu
DiffAb
AbMPNN
RFDiffusion
HyperMPNN
AntiBERTy
AF-Traj
AbGPT
a man with glasses is holding his head in front of a computer monitor
ALT: a man with glasses is holding his head in front of a computer monitor
media.tenor.com
June 5, 2025 at 12:36 PM
It is often very informative to compare many ML methods for structure prediction.

Do all the models converge ?

Are there subtle differences between the models, with some more convincing than others?

Currently testing #alphafold #rosettafold #ESMFold #Omegafold #openfold and #unifold
December 13, 2023 at 9:32 AM
3D structure data. AlphaFold2 and related models, including RoseTTAFold and ESMFold, employ specialized neural network architectures driven by attention mechanisms to infer relationships between sequence and structure. At a fundamental level, these [2/7 of https://arxiv.org/abs/2504.14059v1]
April 22, 2025 at 6:15 AM
significantly outperform those trained on smaller models for both biological concept discovery and contact map prediction. Finally, we present an initial case study demonstrating how our approach enables targeted steering of ESMFold predictions, [5/7 of https://arxiv.org/abs/2503.08764v1]
March 13, 2025 at 6:14 AM
scale SAEs to ESM2-3B, the base model for ESMFold, enabling mechanistic interpretability of protein structure prediction for the first time, and (2) we adapt Matryoshka SAEs for protein language models, which learn hierarchically organized features by [3/7 of https://arxiv.org/abs/2503.08764v1]
March 13, 2025 at 6:14 AM


Balancing Speed and Precision in Protein Folding: A Comparison of AlphaFold2, ESMFold, and OmegaFold

https://www.biorxiv.org/content/10.1101/2025.06.20.660709v1
June 23, 2025 at 12:20 PM
Haven’t seen that so far! I’m doing some side by sides on generated designs versus colabfold and esmfold, but not done yet. Would be curious how the same sequences fold using other models
October 3, 2025 at 5:33 AM
now there's a title that catches the eye.

paper's based on ESMfold predictions. it's an intriguing idea, but i fear the results may be an artifact of the training data. they may be seeing poor generalization of ESM rather than a biological signal
Natural protein structures have evolved exceptional robustness to mutations https://www.biorxiv.org/content/10.1101/2025.08.27.672565v1
September 1, 2025 at 5:21 AM