All questions
Question
Is open-source AI catching up?
Of the 6 moments we cut that answer this, 4 say open models are close behind, one says they’re behind but useful and one says they’re falling behind. Arvind Krishna said on All-In (Oct 21, 2026) that open-weight models can handle most enterprise work at much lower cost, while frontier models remain for high-value tasks. Gavin Baker said on BG2 (Jun 11, 2026) that open models handle cheaper back-office work, while frontier models keep high-value tasks like coding.
Where they stand
Close behind4
Enterprises may reserve frontier models for their hardest tasks
“the open weight models right now are so good. Even if they're three months behind, it's plenty for what we need.”
George KurtzPresident & CEO, CrowdStrike23:28 Open-weight models could handle most enterprise AI at lower cost
“Which enterprise in the world will not accept a 10 times, or in this case, a 40 times cheaper answer for effectively what can do 95% of the work”
Arvind KrishnaChairman & CEO, IBM13:13 Open models now handle most token traffic on Vercel
“I don't need to hire Albert Einstein to solve every problem, like answer every support ticket and spell check my stuff and classify my data.”
Guillermo RauchFounder & CEO, Vercel56:15 Open-source models trail closed ones by about six months
“This is different because we're getting this capability before we've really characterized the shape of the threat.”
John CooganCo-host, TBPN6:21
Behind but useful1
Frontier labs capture the revenue, but open source may carry most tokens
“Frontier might be 90% of the economic value. Open source might be 80% of tokens.”
Gavin BakerManaging partner & CIO, Atreides Management53:34
Falling behind1
Frontier labs are holding back models before public release
“Anthropic had Mythos ready in February. Their model safety cards noted that they have two other models since Mythos that have released.”
Dylan PatelFounder & CEO, SemiAnalysis55:47
Each moment is a few minutes of a long interview, cut by Real Alpha and filed under how it answers by a model that read its claim and quote. Open a moment for its full transcript. Last moment added Oct 23, 2026.