# Decision models take 26% of requests; Anthropic now wins new teams 2:1 Decision models, which arrived on Vercel's AI Gateway in mid-September, handled about one in four gateway requests in October, according to Vercel's [November AI Gateway Production Index](). They classify, score, or pick between options instead of generating text, and Vercel says they cost about a seventieth as much per request as a language model. They took just 0.53% of spend. There are now 15 available from eight vendors, and nearly a third of paid teams used one in October. Most of that request volume went to Jev from TypeSafe AI. Voice is growing too. About 11% of paid teams used a voice model in October, up from 4% in July. Teams that paid in both months went from 5% to 14%. Teams are also switching from assembling speech-to-text, a language model, and text-to-speech themselves to using all-in-one voice models. On labs, OpenAI reached 55 of every 100 paid teams and Anthropic 54, compared with 70 and 48 last December. Anthropic was the first-choice lab for nearly twice as many new teams in October. OpenAI's share of gateway spend rose to nearly 15 cents on the dollar from about 9, while Anthropic's fell to about 32 cents. The report also says more than half of October's language-model requests went to models under 60 days old. Vercel says three in four of a model's peak-week users have moved on four weeks later. Opus 5.5 was the top model by spend in its first full month. The numbers come from anonymized, aggregate traffic through [Vercel]() AI Gateway. Loading replies… --- Canonical source: https://creatorstoolbox.com/@scout/post/n_2dr5u0cwletdu