卖什么
Environments for Long Horizon Work
商业模式拆解
买家：谁掌握预算？
Frontier AI labs and RL data vendors (site: working with four top frontier AI labs, several Fortune 100 companies and leading RL data vendors).
触发点：什么让他们现在购买？
未核实Labs need hard, long-horizon training environments and post-training data for coding, computer use, knowledge work and research agents.
工作流：什么输入变成什么输出？
Aviro designs and builds simulated worlds, converts them into post-training datasets and reinforcement learning environments, and publishes benchmarks (e.g. C4 grounded-reasoning benchmark across 3 simulated company environments).
渠道：怎么接触到客户？
Demo-led ('Book Demo' via cal.com), with benchmarks and research writeups as credibility marketing.
融资历史与报道
| 融资日期 | 轮次 | 金额 | 投资方 | 来源 |
|---|---|---|---|---|
| — | YC 2025 春季 | — | Y Combinator | ycombinator.com |
档案汇总自公开报道和公司官网。商业模式部分是 MaxBC 对这些材料的解读，不是已确认的事实；未知之处留空，不作猜测。 更新于 2026-10-12.