🍬 Superbio Wrapped 2025

Dear Superbio community,

2025 was the year Superbio went from useful to transformative, from an AI app store to a full-fledged Virtual Lab.

Together, we ran real experiments, stress-tested new ideas, and pushed Bio-AI workflows beyond demos and into daily research. From first-time users spinning up protocols in minutes, to power users chaining models in ways we never anticipated, this year was defined by momentum.

So before we look ahead, we wanted to pause and reflect on what we built together.

Here’s Superbio Wrapped 2025 👇

🧪 By the Numbers

• 5,260 model runs executed

• 900+ computational protocols built

• 10,000+ GPU hours orchestrated behind the scenes

• 0 infrastructure setup or coding required 😉

🚀 Top 3 New Features You Loved

1. Virtual Lab (Cloud Jupyter + GPUs)
   End-to-end Bio-AI pipelines, reproducible and runnable in minutes.

2. Plain-English Pipeline Builder 
   Describe the experiment, Superbio assembles the workflow.

3. Public Superbio APIs
   Access and execute 650+ models on Superbio using Python.

🧬 Top 3 New Models Added

From generative binder design to structure-aware sequence optimization, this year we expanded our catalog with tools researchers actually use:

• Binder Design with Boltz-1 — end-to-end binder design pipeline, scaling to 500,000 binders with ease

• LigandMPNN — structure-aware sequence design for protein–ligand and protein–protein interfaces

• Boltz-2 — joint modeling of complex biomolecular assemblies and binding affinities

Coming soon: BoltzGen — large-scale generative binder design, built for speed and scale

(What a year for Boltz developers!)

🧠 Most Common Use Cases

Here’s what researchers used Superbio for most in 2025:

• Going from target to real drug candidates
  Starting with a biological target and generating binders or variants to explore

• Scaling exploration before the wet lab
  Designing and filtering hundreds of thousands of candidates in silico

• Answering “what should we test next?”
  Ranking and narrowing large design sets to a focused shortlist

• Re-running and sharing experiments
  Saving protocols so results are reproducible and easy to revisit

🧑‍🔬 Community Signals We Loved

• Most-requested feature:
  An LLM in the loop for interacting with Jupyter notebooks (coming soon 👀)

• Most common “aha” moment:
  Realizing pipelines can be reused, shared, and rerun — not just executed once

• Most common user question:
  “Can I bring my own models?”
  (Yes. Always yes.)

🎓 Top 3 Webinars of 2025

• Programming Biology in English — a first look inside Superbio’s Virtual Lab (recording coming soon)

• Scientific Agents for Biotech R&D — where we introduced the Superbio Pipeline Builder

• Blazing a Trail to Universal Vaccines — a conversation with Dr. Jacob Glanville on the future of vaccines

🔮 Looking Ahead

In 2026, we’re doubling down on:

• Enabling full autonomy for the Superbio Co-pilot to execute protocols and help scientists interpret results

• Expanding high-quality, production-ready models

• Model fine-tuning and head-to-head benchmarking

Thank you for building, experimenting, and pushing Superbio forward with us.

Wishing you happy holidays,
The Superbio Team 🩷