Teams running coding agents get a flat $30 plan with $80 of credit on GLM 5.3 and Kimi K3 through AI Gateway, with zero data retention, and requests stop at the cap with no overage charges.
Key takeaways · AI-distilled
Vercel names fx, Command Code, and Pi as coding agents that work with tokenThe chunk of text a model reads and writes in — roughly three-quarters of a word — and the unit AI usage is billed in.Full definition → Pass, and says teams can also build their own AI agentAn AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.Full definition → with the AI SDK and switch between the included models on the shared team credit.
Zero Data Retention is on by default: requests run on selected providers with inferenceRunning a trained model to get answers — the phase where AI is actually used, as opposed to trained.Full definition → in the US and EU, prompts and responses are deleted after each request, and prompts are not used to train models, per the post.
The subscription renews monthly unless canceled, and any unused credit does not roll over into the next billing period.
Setup runs through the Token Pass page in the team's AI Gateway dashboard: activate the subscription, create a dedicated Token Pass API key, connect the agent, then select GLM 5.3 or Kimi K3.
Terms in this piece · Glossary
token — The chunk of text a model reads and writes in — roughly three-quarters of a word — and the unit AI usage is billed in.
AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
inference — Running a trained model to get answers — the phase where AI is actually used, as opposed to trained.
