Anew Labs

Fight Disease, Advance Human Health

[Anew Labs @AnewOffcial]

Thrilled to introduce AnewDDE, an agentic drug discovery engine that brings together biomolecular structure modelling, molecular design, affinity prediction, and scientific reasoning in a closed-loop workflow.

Read the technical report: https://t.co/ygaSR3s7QR

[Anew Labs @AnewOffcial]

https://t.co/TJh2UGlCXL

[Anew Labs @AnewOffcial]

Anew Labs has been named to the 2026 Endpoints 11, @endpts' annual list of biotech's most exciting startups.
We drive breakthroughs in AI to tackle longstanding challenges in drug discovery.
Thank you to @endpts and @AndrewE_Dunn , and congratulations to our fellow honorees.

[Anew Labs @AnewOffcial]

AnewMind: a 100B-scale LLM post-trained for pharmaceutical reasoning.

Top-5 zero-shot on all four modalities. #1 on cyclic-peptide PAMPA permeability (0.737).

On PharmBench — our 50-scenario expert benchmark — it holds its own against frontier models.

[Anew Labs @AnewOffcial]

One nanobody campaign, run by AnewDesign as a closed loop:

3,092 candidates → 287 clusters → 50 sent to the bench
7/50 under 400 nM by SPR
150 clones total → 16 at single-digit nM

Wet-lab results feed the next design round, not a final verdict.

[Anew Labs @AnewOffcial]

AnewAffinity ranks binding fast enough to steer a design cycle: ~1.5s per ligand-pair edge on one GPU.

On 8 retrospective JACS systems (199 ligands): Spearman 0.724 vs 0.620 for local Boltz-2.

And every estimate exposes the windows behind it.

[Anew Labs @AnewOffcial]

AnewFold, the structure module of AnewDDE.

76.2% top-1 on FoldBench-AB. 77.4% on similarity-filtered FoldBench protein–ligand.

On a post-cutoff molecular-glue benchmark: 71.8%, vs 45.0% (Protenix-v1) and 29.8% (Protenix-v2).

Hardest cases, widest margin.

[Anew Labs @AnewOffcial]

Learn more about Anew Labs: https://t.co/rhFBGGMrGV

[Anew Labs @AnewOffcial]

Thrilled to introduce AnewDDE, an agentic drug discovery engine that brings together biomolecular structure modelling, molecular design, affinity prediction, and scientific reasoning in a closed-loop workflow.

Read the technical report: https://t.co/ygaSR3s7QR