Samaipata published MDIA v1.0.62, which delivers 65.1% using the same length-adjusted scoring procedure as OpenAI.
Public source
Publisher name
Public post
OpenAI’s latest medical AI results look familiar. GPT-6 Astra reports 63.4% on HealthBench Professional after adjusting for response length. Its highest-scoring configur…
Company
Samaipata
We invest in tech companies with network effects at the early stage across Europe.
- Industry
- Venture Capital and Private Equity Principals
- Location
- Europe, GB
- Company size
- 11–50 employees
About Samaipata
Samaipata is an early-stage founders’ fund investing in tech companies with network effects across Europe and beyond. The fund was founded in 2016 by two entrepreneurs, José del Barrio and Eduardo Díez-Hochleitner. José was the co-founder and CEO of the leading food-delivery platform in Spain (La Nevera Roja) which was sold for $100 million. Eduardo was founder of a 3D sound system sold to Dolby (IMM Sound), former Partner at Apax Partners, a prolific business angel, and the current Chairman at MásMóvil, recently valued at €20 billion, becoming the first Southern European Decacorn. If you’re an early-stage founder aiming at changing the world through innovation, get in touch with us! >> https://samaipata.typeform.com/to/k43fN6?utm_source=samaipata-linkedin
See moreLatest activity
Latest activity from Samaipata
6 signals
Research & Knowledge
Samaipata published MDIA v1.0.62, which delivers over 71% without the length adjustment compared with Astra’s reported 69.5%.
Research & Knowledge
Samaipata published MDIA v1.0.62, which delivers $0.05 per prompt.
Technology & Infrastructure
Samaipata introduced two systems using TietAI to handle telehealth calls for patients.
Discover more
Similar signals
Similar public activity from other companies.
Research & Knowledge
May Ventures
May Ventures announced that OpenAI announced its new model Astra, which is rated Critical for cybersecurity and features recurrent latent space reasoning.
Research & Knowledge
Brookstreet Equity Partners LLP
Brookstreet Equity Partners LLP reported that Ox Alpha appeared anonymously and immediately caught developers' attention in coding tests, with 80% vs Claude Fable 5 at 65% and GPT-5.6 Sol at 52%, while using roughly 3× fewer output tokens in one bug-fixing comparison.
Products & Services
Village Global
Village Global's Wispr Flow model is now 3x more accurate than OpenAI's models and growing 40% month over month.
Research & Knowledge
Zero Prime Ventures
Zero Prime Ventures published a write-up detailing that GPT-5.6 and Agno's Learning Machines scored 100 on the public set across 183 levels in 25 games.
Research & Knowledge
Heartcore Capital