Insitro Opens Lilly TuneLab In Vivo PK Models to External Biotech Teams
On 6 October 2026 insitro said Lilly TuneLab now offers biotech users the in vivo pharmacokinetic models the partners trained on decades of Lilly preclinical measurements.
What changed
On 6 October 2026, insitro announced that machine learning models it built with Eli Lilly to predict in vitro and in vivo small molecule behavior are available to select biotech companies inside Lilly TuneLab. The models draw on multi species preclinical data Lilly generated across hundreds of thousands of unique molecules over decades, covering absorption, distribution, metabolism, excretion, and toxicity style measurements.
insitro stated the models will continue updating as the underlying dataset grows. The release follows a collaboration announced in September 2025 pairing insitro compute with Lilly proprietary ADMET scale data.
Why it matters
Most biotech teams still extrapolate human dose predictions from sparse in vitro panels because they lack in vivo scale history. Access to Lilly trained in vivo pharmacokinetic models, even through a gated portal, shifts early portfolio decisions: chemists can prioritize series with credible exposure forecasts before committing large animal studies.
For pharma platform strategists, the move is another example of monetizing legacy measurement archives through AI wrappers rather than hoarding them inside one pipeline.
Who is affected
Small and mid biotech medicinal chemistry leads, CRO planners, Lilly partnership scouts, and competitors such as Schrödinger or Relay selling predictive ADMET suites.
What to do next
If you qualify for TuneLab access, benchmark the in vivo PK model against three internal compounds with known clinical exposure. Document where the model beats your legacy rules and where it fails before trusting it for series triage.
What to watch
insitro disclosures on TherML platform integrations and any Lilly TuneLab expansion beyond PK into safety or formulation endpoints.
Sources
- Primary. insitro, Machine learning models for in vivo pharmacokinetic prediction now available in Lilly TuneLab (6 October 2026). Describes model scope, data provenance, and TuneLab availability.
- Secondary. Lilly, TuneLab platform overview (accessed 12 October 2026). Confirms portal exists; detailed gating requires partner login.