If OpenAI's next model can't beat two-month-old Opus 5.5, it may be losing the race
OddConsideration7290 · reddit · 2026-11-15
Building on METR's report on Anthropic's acceleration, the author argues OpenAI's position is genuinely unknowable since it hasn't shipped a new model.
Three scenarios: if OpenAI's latest still trails Anthropic's best, it's losing; matching Opus 5.5 is still bad news since that model is nearly two months old; only reaching or beating Fable 5.2 keeps the race alive. The author calls this their most anticipated release ever and jokes about going all-in on Anthropic's IPO if it disappoints. They're mildly optimistic — liking GPT-6 Astra and GPT-6.1 Sol and OpenAI's current Pareto frontier — but note OpenAI's models are painfully slow and need to ship something both strong and fast.
More from Companies & People
- NY Post exposé details Jensen Huang's office tantrums after his ouster from Nvidia — leveragedupside · 2026-11-15
- Nvidia ouster reignites debate over the value of founder-led companies — JOBhakdi · 2026-11-15
- Grok bot engineer's AIE talk reveals how the team ships fast without cutting quality — gaganghotra_ · 2026-11-15
- AI Tinkerers hosted nearly 300 events in 90 days with 500+ project demos — gabe_ragland · 2026-11-15
- "Jensen played GOD": founder accuses Nvidia of circular deals, smuggling and a $5T bubble — bindureddy · 2026-11-15
- $280k GTM jobs ask for Python (71%) far more than Google Analytics (20%) — VibeMarketer_ · 2026-11-15