Deep Dive: How Many Unreleased Advanced Models Do AI Labs Actually Have?
Concern-Excellent · reddit · 2026-08-10
Based on recent model leaks (such as OpenAI's Astra and Doug), this long post speculates deeply on the internal development timelines and model reserves of top AI labs.
- Timeline Breakdown: The author notes that pretraining takes about 6 months, post-training about 1 month, while internal testing, guardrail deployment, external verification, and government approvals take up even more time. For instance, Claude Mythos was developed in February, released to partners in May, and went public in July.
- Generational Buffer: It is widely believed that top companies keep a 1-2 generation buffer internally. This means the models the public sees today were actually finalized six or more months ago.
- Internal AGI Perception: Because companies may already possess models far exceeding public releases (like the rumored Doug), researchers inside these labs feel the approach of AGI and technological advancement much more intensely than outside observers.
More from AGI Musings
- Personal shopping agents bumped her secondhand spending from $365 to $820 in 2 months — davidyin44 · 2026-10-16
- There's no formal route to frontier AI research: 19-year-olds are advancing tech from their basements — repligate · 2026-10-16
- 'Will you end humanity?' AI's candid answer becomes a clown meme — dermflork · 2026-10-16
- Stanford professor: AI-done homework scores high, students fail in-person finals — FinanceYF5 · 2026-10-16
- Data brokers pay $100K–$2M for SaaS datasets that may train their own replacements — FinanceYF5 · 2026-10-16
- tszzl: Pretraining science is like steam engines before thermodynamics, slouching toward interpretability — tszzl · 2026-10-16