Benchmark
Measure the gain on target discovery.
Compare the full Helical workflow with biological foundation models alone, generalist AI tools and a differential expression baseline.
Recovery@100 vs. Open Targets top-5% evidence, averaged across 7 indications. Reasoning Models = avg of Codex + Kimi K2.7. Bio Foundation Models = Fine-tuned Geneformer. Higher is better. Retrospective target recovery. Experimental validation is a separate step.
Click here to see the full benchmarkHow it works
Make biological foundation models work on your data.
Target identification and biomarker discovery.
01
Find the models that work on your data.
Evaluate biological foundation models against your scientific questions. Compare results on held-out data before choosing which models to use.
02
Teach them your biology.
Post-train models on your disease data and experiments. Measure what improves against the criteria your team has defined.
03
Run target and biomarker workflows.
Combine your evaluated models with scientific tools to run virtual experiments. Compare predictions and rank target and biomarker candidates for follow-up.
04
Decide which candidates to test.
Assess candidates against orthogonal evidence from genetics, experiments and the literature. Inspect the proposed mechanisms, then use your experimental results to refine the next round.
Trusted by
Trusted by the world's leading biopharma R&D teams.
Your biological intelligence
Built for research teams.
SOC 2 TYPE II (in progress)The models you adapt stay yours.
Build on your proprietary data and retain the model weights learned from it. Evaluate new models as your scientific needs evolve.
Deployed Scientists.
Our Deployed Scientists work with your team to define evaluations, adapt models and assess the evidence for your research programmes.
One platform, from models to decisions.
Bring model evaluation, post-training and target and biomarker workflows together in one virtual lab, connected to your research environment.
Better target and biomarker decisions.
Build biological intelligence around your data and research programmes.