Mostly True
Anthropic and OpenAI, both of which previously voiced support for slowing the pace of AI development, have rolled out new AI models emphasizing speed and cost efficiency, drawing attention to a gap between their regulatory rhetoric and their commercial strategies. The overall claim is judged to be largely consistent with the facts, though some nuance applies.
Both companies have at times publicly warned about the risks of racing ahead in AI development, with Anthropic in particular having been associated with calls to pace frontier AI progress. Yet with their latest model releases, the two firms have emphasized faster performance and lower per-use costs — attributes aimed squarely at winning developers and enterprise customers in a price- and speed-sensitive market.
The new models are being marketed on efficiency: delivering competitive capability at reduced cost and higher speed. This places both companies in direct competition on "value for money," a notable shift from the earlier framing of slowing down frontier development. The commercial logic is straightforward — faster, cheaper models expand usage and revenue even as safety rhetoric continues.
The contrast does not mean the companies have abandoned safety positions; both continue to publish safety frameworks and research alongside product launches. But the simultaneous push for speed and cost efficiency does sit uneasily with earlier calls to slow the frontier, and observers have pointed to this tension.
Anthropic and OpenAI did urge restraint at the frontier — and both are now racing on speed and cost with their newest models. On balance, the claim is Mostly True.
Verdict: Mostly True