The Copenhagen company says its Symphony model has outscored OpenAI on HealthBench Professional, and is offering credits and regulatory help to founders building healthcare AI worldwide.
Corti, the Copenhagen-based clinical AI company, has launched a no-equity accelerator for healthcare and life sciences startups, opening its Symphony model stack to founders worldwide at a moment when the regulatory cost of building medical AI in Europe has rarely been higher.
The company said on Tuesday that Symphony, its flagship clinical-grade model, has outscored OpenAI on HealthBench Professional, the benchmark for realistic clinician conversations that OpenAI released last month alongside its ChatGPT for Clinicians product.
The new Startup Acceleration Program offers up to $5,000 in credits across the full Symphony stack, which spans agents, medical coding, speech-to-text, and text generation and is, according to the company, trained on more than 1.5 million hours of clinical audio.
The 💜 of EU tech
The latest rumblings from the EU tech scene, a story from our wise ol’ founder Boris, and some questionable AI art. It’s free, every week, in your inbox. Sign up now!
Participants also receive direct time with Corti’s clinical and regulatory team on EU AI Act, MDR, and data-residency questions, founder-led roadmap webinars, and invitations to Corti events in New York, Copenhagen, London, and Berlin.
Applications are open today on a rolling basis with a one-week turnaround. There is no pitch process and no equity component, and the programme is open to pre-seed through Series B companies in healthcare, clinical workflows, or adjacent life sciences.
The launch sits inside a sharpening regulatory backdrop. In April, OpenAI rolled out free clinical AI to every verified American physician, nurse practitioner, physician assistant, and pharmacist.
A week later, OpenEvidence, the clinical AI search platform valued at $12 billion and used daily by roughly 40% of US physicians, withdrew from the UK and the EU, citing regulatory uncertainty under the EU AI Act.
The high-risk-system obligations behind that decision begin to apply on 2 August 2026, although AI embedded in CE-marked medical devices regulated under MDR or IVDR has a separate Article 6(1) timeline currently set for 2 August 2027.
Europe’s structural cost layer is the other half of Corti’s argument. EU MDR certification alone costs founders between €200,000 and €600,000 per device and takes 12 to 18 months, according to industry navigation guides.
Galen Growth’s Q1 2026 analysis shows European digital health funding consolidating in larger, later-stage rounds, with investors increasingly favouring companies that can show clinical evidence and workflow integration rather than headline performance.
“The future of healthcare AI won’t be built by one company. It will be built by thousands of teams, each with deep knowledge of a specific care setting, workflow, or patient population,” said Andreas Cleve, Corti’s co-founder and chief executive.
“Our job is to give those builders a head start: the leading clinical AI model, the evidence base behind it, and a path to production we’ve already navigated for regulated health systems. So they can focus on what only they can do, the workflow, the patient population, the problem they actually understand.”
Among the development teams already building on Corti is Aisel Health, a European startup focused on psychiatric workflows.
“Psychiatrists are a scarce and highly specialised resource,” said Augusta Klingsten Peytz, Aisel’s co-founder and chief executive. “
They should be focused on one thing only: making clinical decisions, everything else needs to go. Yet today, the majority of a psychiatrist’s time is spent not on clinical decision-making, but on the administrative and repetitive workflows surrounding it.
By using Corti, we at Aisel can focus on delivering specialised psychiatric workflows that help clinicians regain capacity, rather than rebuilding the clinical-grade foundation underneath.”
Corti has raised $100 million to date, with offices in Copenhagen, New York, and London. Its Symphony for Medical Coding release in April claimed a 25% accuracy edge over OpenAI and Anthropic on ACI-BENCH and MDACE, two academic medical-coding benchmarks, and the company says the stack powers AI for systems serving more than 100 million patients annually, including the NHS.
The programme’s premise, in Corti’s framing, is that the conditions making Europe difficult for horizontal model providers are the conditions vertical clinical-AI players were built for. Applications are open from today.
