Open-weight models now handle 71.5% of AI Gateway tokens and 23% of spend

Open weight: money follows tokens

Vercel says AI Gateway routes tens of trillions of tokens each month between production applications and AI labs, and it publishes a monthly index based on that traffic. In August, open-weight models handled the majority of token volume for the first time. According to the October report, spending is now catching up: open weight's share of spend rose from one in twenty gateway dollars in June to nearly one in four in September. September also brought the largest monthly gains in both open-weight token share and spend share in the index's history.

The growth comes from both ends of the price range. Cheap models drive token volume, while expensive ones drive spend. DeepSeek-v4.1-Flash alone processed nearly half of all September tokens but took just 2.6% of spend, at a price per token about a fifth of the open-weight average. Kimi K3, by contrast, accounted for nearly half the increase in open-weight spending while processing far fewer tokens. Models priced at least four times the open-weight average made up about two-thirds of that increase.

Frontier spending is also rising. Since December, spend on frontier models has grown tenfold, while spend on open-weight models has grown more than seventyfold. In September, 68% of teams that put more money into open-weight models did the same for closed-weight models. More than two in five teams already paying for frontier models spent more on them than in August, the highest share this year.

Bigger, cheaper and less loyal

September's gateway spend was about 13 times December's, an increase of 1,180%, while the average price per token fell 47%. According to Vercel, the median paid customer active in both months more than doubled its spend.

Loyalty to individual models and labs is weakening. In September, nearly one in two teams changed their most-used model, and almost one in three changed their most-used lab, an index record. Both measures rose for a fifth consecutive month. Teams also use many models: since December, those processing at least 10 million tokens used an average of 15 different models, rising to 26 for teams at 100 million tokens and 42 for those at a billion. Vercel concludes that the model is becoming a replaceable part of a much larger investment.

Chinese models, sold by American providers

More than two-thirds of September's gateway spend on Chinese-origin models went to US-based providers, up from less than a tenth in December. About $2.50 was spent on Chinese models through US-based providers for every $1 spent through China-based ones.

• DeepSeek models processed 58% of all gateway tokens for 6.6% of spend. More than four in five of those tokens were served by DeepSeek itself, yet two in five dollars spent on its models went through US-based providers.
• Moonshot AI, the lab behind Kimi, accounted for 3.6% of tokens and 13.2% of spend. US-based providers served nine in ten of those tokens.
• Z.ai, creator of GLM, accounted for 7.2% of tokens and 4.5% of spend. US-based providers served seven in ten.

September launches: Jev, Opus 5.5 and GPT-6

Jev, from TypeSafe AI, was the fastest-adopted model in AI Gateway history, reaching 35% of paid teams in its first two weeks, nearly double the reach of second-placed Claude Sonnet 5.5. Of the teams that tried it in week one, 55.6% returned in week two. Vercel describes Jev as a decision model: its requests are faster, cheaper and use a fraction of the tokens of typical language-model requests. It took just 0.27% of gateway spend in its first two weeks.

Claude Opus 5.5 took 11.03% of spend, more than twice the share of any other September launch, followed by GPT-6 Astra at 4.86% and DeepSeek V4.1 Flash at 4.19%. Almost nine in ten dollars for Opus 5.5 came from teams already using Opus 5. Meanwhile, Anthropic lost nearly 20 percentage points of spend share in September, with Opus accounting for 17 of them. Its overall share fell from 63.7% to 43.9%, below half of total gateway spend for the first time.

What it means

The index suggests that spending on frontier labs is still growing, but spending on alternatives is growing faster. Vercel notes that a company can pay a frontier lab more while relying on it for fewer workloads. For labs, that means winning business model by model, since teams are switching more often and spreading usage across many models.