1.4 Trillion Tokens: Alibaba Qwen3.6-Plus Breaks OpenRouter Daily-Usage Record

Record set within 24 hours

Alibaba (阿里) said to have its new model Qwen3.6-Plus (千问 Qwen3.6-Plus) reach a major usage milestone almost immediately after release. It has been reported that OpenRouter posted on X that the model — launched April 2 — processed more than 1.4 trillion tokens in a single day, breaking the platform’s record for single‑model daily calls. The announcement frames Qwen3.6-Plus as the first model on OpenRouter to exceed the trillion‑token mark in one day and the strongest performer among models released in 2026, reportedly based on developer and enterprise token consumption statistics.

Why the platform matters

OpenRouter is described by IT Home as the world’s largest AI model API aggregation platform, one that pools traffic for models such as Claude, GPT and Qwen and publishes leaderboards based on token usage. Because the platform tallies real developer and enterprise consumption, its numbers are often read as a proxy for commercial traction and third‑party adoption — though platform‑level claims should be treated as reported rather than independently verified.

Model performance and broader implications

Qwen3.6-Plus was touted by Alibaba for breakthroughs in programming and agent capabilities; it has been reported that the preview ranked first among Chinese models in the programming subleaderboard of the international Arena benchmark, helping Alibaba secure a No. 2 position among AI institutions on that metric. The surge on OpenRouter underscores how quickly Chinese models can attract heavy usage once made available to developers.

A sign of rising competition

This milestone arrives against a backdrop of intensified U.S.–China technology rivalry and export controls that have shifted how both sides deploy and commercialize advanced AI. Will Western providers respond with new pricing, partnerships or expanded hosting to stem token outflow? For now, the OpenRouter figures offer a striking snapshot of adoption — and a reminder that model rankings are increasingly part technical merit, part commercial distribution. It has been reported that OpenRouter’s post and IT Home’s coverage are the primary sources for these claims.