#Nextflow
🌱 Build your skills progressively through beginner-friendly courses:

→ Hello Nextflow - Learn to develop your own Nextflow pipelines

→ Nextflow Run - Learn to run and configure existing Nextflow pipelines

→ Nextflow for Science - Apply learnings to genomics, RNA-Seq, or imaging
November 16, 2026 at 9:18 AM
Nextflow Training Week by @seqera.io kicks off today, and it’s not too late to join! 🔗 Online and free to attend: hubs.la/Q04FkyxJ0
November 16, 2026 at 9:18 AM
Join us Nov 18 to recap the latest Nextflow release and what it means for your pipelines, from pipeline composition, Global Cache and CLI type checking to reusable subworkflows via the Nextflow Registry and Agentic Nextflow. Free to join: hubs.la/Q04D_DBR0
November 11, 2026 at 4:13 PM
The final module of our 'Byte-sized green computing' course will occur on 30 November 2026. Join Josua Carl and Nadja Volkmann for 'nf-co2footprint: estimating the carbon footprint of Nextflow pipelines'.

Registration is free at this virtual course: www.ebi.ac.uk/training/eve...

📊 🖥️ 🧬
November 16, 2026 at 9:01 AM
Nextflow Frontier Introduces Innovative Discovery System for Complex Scientific Challenges#United_States#AI#New_York#Nextflow#Discovery_System
Nextflow Frontier Introduces Innovative Discovery System for Complex Scientific Challenges
Nextflow Frontier has launched a groundbreaking discovery system aimed at solving complex scientific issues, outperforming existing AI models in key benchmarks.
third-news.com
November 12, 2026 at 2:47 PM
The difference came after the first good answer. Rosalind and Claude Science stopped there. Nextflow Frontier kept looking for a better one.

Median scored attempts per task: 506, against 56 for Rosalind and 3 for Claude Science.
November 12, 2026 at 3:07 PM
Rosalind and Claude Science also had to find and run their own tools. Nextflow Frontier separates the science from the compute. Each idea becomes a Nextflow pipeline that runs at scale on @seqera.io Platform, and every score feeds the next round.
November 12, 2026 at 3:00 PM
On average, Nextflow Frontier beat 96% of the prior methods on each task's public leaderboard. Rosalind beat 75%. Claude Science beat 66%.

It ranked highest in all four areas: drug discovery, protein engineering, single-cell omics and biomedical imaging.
November 12, 2026 at 3:07 PM
Happening next week!📆 Don't miss your chance to earn an official @nextflow.io training certificate, from the developers of Nextflow. We hope to see you there: hubs.la/Q04DKQ7V0
November 9, 2026 at 12:17 PM
🧭 Explore @nf-co.re best practices and advanced development techniques:

→ Hello nf-core - Learn to utilize code and best practices from nf-core

→ Side Quests - Deep dive into standalone advanced Nextflow topics

→ Training Collections - Apply Side Quests via recommended learning paths
November 16, 2026 at 9:18 AM
How do you beat both OpenAI's Rosalind and Anthropic's Claude Science on biology problems?

We gave Nextflow Frontier and both of them the same 22 BioML-bench tasks, the same data and the freedom to keep experimenting as long as they liked. Here's what happened.🧵
November 12, 2026 at 3:07 PM
Stopping at "it works" can be good engineering. In science, you don't know where the best answer is until you've looked.

Full results, task by task: hubs.la/Q04F5jQ10
Nextflow Frontier, Rosalind and Claude Science on BioML-bench
On the same 22 BioML-bench tasks, Nextflow Frontier's mean leaderboard percentile was 96, against 75 for OpenAI's Rosalind and 66 for Anthropic's Claude Science.
hubs.la
November 12, 2026 at 3:07 PM
Stopping at "it works" can be good engineering. In science, you don't know where the best answer is until you've looked.

Full results, task by task: hubs.la/Q04F5jQ10
Nextflow Frontier, Rosalind and Claude Science on BioML-bench
On the same 22 BioML-bench tasks, Nextflow Frontier's mean leaderboard percentile was 96, against 75 for OpenAI's Rosalind and 66 for Anthropic's Claude Science.
hubs.la
November 12, 2026 at 3:00 PM
🌱 New to Nextflow? Build your skills progressively with beginner courses including Hello Nextflow, Nextflow Run & Nextflow for Science

🧭 Already familiar? Explore @nf-co.re best practices and advanced pipeline development techniques with Hello SideQuests & Training Collections
November 9, 2026 at 12:10 PM
Nextflow Frontier uses language models for individual research steps. Its Adaptive Discovery Model decides what to try next, balancing new ideas against refining the best so far, an approach designed for biology's graded results.
November 12, 2026 at 3:07 PM
Nextflow Frontier uses language models for individual research steps. Its Adaptive Discovery Model decides what to try next, balancing new ideas against refining the best so far, an approach designed for biology's graded results.
November 12, 2026 at 3:00 PM
On average, Nextflow Frontier beat 96% of the prior methods on each task's public leaderboard. Rosalind beat 75%. Claude Science beat 66%.

It ranked highest in all four areas: drug discovery, protein engineering, single-cell omics and biomedical imaging.
November 12, 2026 at 3:00 PM
The difference came after the first good answer. Rosalind and Claude Science stopped there. Nextflow Frontier kept looking for a better one.

Median scored attempts per task: 506, against 56 for Rosalind and 3 for Claude Science.
November 12, 2026 at 3:00 PM
Take EGFR, a protein behind some lung cancers. Rosalind reached 0.864 in 51 attempts, then decided further changes were marginal.

Nextflow Frontier changed how it described each molecule over 820 attempts and reached 0.887, surpassing the previous leaderboard best.
November 12, 2026 at 3:07 PM
Take EGFR, a protein behind some lung cancers. Rosalind reached 0.864 in 51 attempts, then decided further changes were marginal.

Nextflow Frontier changed how it described each molecule over 820 attempts and reached 0.887, surpassing the previous leaderboard best.
November 12, 2026 at 3:00 PM
Full write-up on how Nextflow Frontier got there, including the analysis behind the hypothesis and the benchmark it was scored against: hubs.la/Q04DXz7d0
Ranking molecules for a cancer drug target
Nextflow Frontier ranked 144 molecules for EGFR binding with two small chemistry models, scoring 0.887 on PR-AUC against 0.866 for a billion-parameter network.
hubs.la
November 10, 2026 at 8:54 PM
Rosalind and Claude Science also had to find and run their own tools. Nextflow Frontier separates the science from the compute. Each idea becomes a @nextflow.io pipeline that runs at scale on Seqera's platform, and every score feeds the next round.
November 12, 2026 at 3:07 PM
How do you beat both OpenAI's Rosalind and Anthropic's Claude Science on biology problems?

We gave Nextflow Frontier and both of them the same 22 BioML-bench tasks, the same data and the freedom to keep experimenting as long as they liked. Here's what happened.🧵
November 12, 2026 at 3:00 PM
💡Early on, Nextflow Frontier ran an analysis that led to a hypothesis: the signal might be in what a molecule’s atoms can do, not just how they’re connected.
November 10, 2026 at 8:54 PM
The result is ahead of every published result on this task, even beating specialized billion-parameter models.

Full write-up on how Nextflow Frontier got there, including the analysis behind the hypothesis and the benchmark it was scored against: hubs.la/Q04DXz7d0
Ranking molecules for a cancer drug target
Nextflow Frontier ranked 144 molecules for EGFR binding with two small chemistry models, scoring 0.887 on PR-AUC against 0.866 for a billion-parameter network.
hubs.la
November 10, 2026 at 9:00 PM