Stanford Rejuvenation A.I. Benchmark leaked - FINALLY someone pushing A.I. corps to take aging seriously
The Stanford Rejuvenation A.I. Benchmark has reportedly leaked, aiming to encourage A.I. companies to prioritize aging science. This benchmark utilizes real cells and organisms in a biosafety lab for its final competition each season. The leak is seen as an opportunity to push A.I. corporations to take aging seriously, as they are motivated by competitive success. The hope is that by publicizing scores, A.I. companies will be incentivized to accelerate research into aging science.
Time & source
Times shown in UTC
Display time zone: UTC
Local time zone unavailable; showing UTC.
PublishedOffset at this time: UTC+0Sep 29, 2026, 01:06 UTC
IngestedOffset at this time: UTC+0Sep 29, 2026, 21:00 UTC
- Published
- Sep 29, 2026, 01:06
- Ingested
- Sep 29, 2026, 21:00
- Source type
- Dev community
- Tier
- Community
- Source status
- Healthy
Tier is a per-source editorial setting, not a per-item score.
It actually uses real cells/organisms in a biosafety lab for the final battle each season. I found it in the science category but I think they're still in stealth. Full results
Claude Opus refused everythying bio related but did really well when it didn't. The best open source models actually did well. there's a LOT of room for model improvement.
Thought I'd spread the word (early hah) so that we can push the A.I. companies to prioritize aging science. They only care if they can win at something, well now they have their battle arena.
That OpenAI researcher @ MajmudarAdam wrote: "the things the models are really bad at, of which there are still many, are things that they have not been trained on. maybe there are the things the models can/will never be trained on"
In other words, if they don't focus training on it, the A.I. models will suck at it.
I do wish the benchmark researchers would publicize these results more. Maybe that's where the community comes in?
Seems unfortunate for A.I. companies to spend millions on compute solving theoretical math problems for marketing while people are dying in droves. They should at least improve medical research and aging science in parallel with math.
Maybe re-tweeting these scores @ them will push A.I. companies to try harder. I'm sure people can figure aging science out eventually, but doing it much faster with A.I. help would be great.