asi1-mini: the speed tier
asi1-mini builds on Qwen3, the open-weight model series from the Qwen team at Alibaba Cloud, released under the Apache 2.0 license. We tune a Qwen3 checkpoint for speed: fast, simple tasks like summaries, classifications, and real-time replies - the workhorse tier that keeps high-volume workflows fast.
Intelligence
Speed-optimized
asi1-mini is our speed tier: it accepts a 262K context window and is tuned for latency-sensitive work - real-time chat, autocomplete, classification, and routing - rather than maximum benchmark depth.
Cost of outcome
Lowest in the family
For read-large/respond-small workloads, asi1-mini is the cheapest way to run the asi1 family. We benchmark latency and quality across the fast-model field and tune the checkpoint for the high-volume workloads our users run every day.
What it does best
- Real-time chat and voice replies
- Classification and autocomplete at volume
- A 262K context window for large documents and long histories
- The lowest cost per task in the asi1 tier set
We do the benchmarking. You get the outcome.
Part of how ASI:One works: we benchmark every model against the tasks your team actually runs - long tool chains, document analysis, code review - before it ships behind a tier. Then we work with you to tune and train the model to your requirements, so you never spend time researching models, comparing options, or upgrading and swapping them out as the market moves.
Try it in ASI:WorkVerify the latest scores on Qwen3 on GitHub.