Persimmon v0.1 Ultra doubles Multi-User Turing Test score from 18% to 48%

gharik · x · 2026-11-21

humansand released Persimmon v0.1 Pro and Ultra, two new tiers using test-time scaling to make every conversation turn more realistic, alongside its first widely available user-model API — the team says it spent significant time on release safety.

Key result: Ultra more than doubles Persimmon's score on the 8-turn Multi-User Turing Test, from 18% to 48%. The 50% mark means indistinguishable from real humans under a held-out judge, making this one of the strongest publicly reported results in AI human-impersonation to date.

Related event: Humansand Launches Persimmon v0.1 Pro/Ultra, Doubling Turing Test Scores(8 posts)→

Original post →

More from Models

Models channel →