AI Agent StoreSqwish · Native API verified
kettei core · Sqwish API
Route a request, check a record or choose an action from answers you define. Use your own labeled examples to choose the smallest model that works well enough.
Released 2026-10-14 · First reviewed 2026-10-25 · Checked 2026-11-03.
We still need comparable independent tests. You can compare the confirmed features and limits below, but we cannot yet give this version an Agent Fit or Practical Value rating. Test it on records with known answers, including cases that should go to a person.
What you can use today
Token prices measure the small pieces of text the model reads or produces. Capacity is the amount the service accepts in a request; check the limits below for any fallback to another model. A dash means the rate is unverified.
Exact provider API ID: kettei-core. Accepts text; this model does not read images. Automatic fallback is on by default and can choose another base model. Use fallback: none for an exact-model test, and record the returned model and version.
Check account access and model readiness before relying on it. Publisher benchmark and speed claims are not independent results here; downloadable core weights are not yet verified. Optional zero data retention for inference requires an organization agreement and is off by default. It excludes stored usage records and training data.
Up to 65,536 billed input tokens and 128 questions per base-model request. Fine-tuned models have a separate 32-question limit.
Output: Yes/no probabilities, choices and scores; no generated explanation. Compare decision models, limits and input costs ↗
Base input costs $0.02 per million tokens; output is free. Fine-tuned rates differ. A fallback is billed at the serving rate, capped at the requested model’s rate. Include retries when comparing the total cost.
Provider release and specifications ↗
Check availability and plans ↗
More setup and licensing details ↗
More setup and licensing details ↗
More setup and licensing details ↗
Test it before switching
Give each model the same examples, tools and definition of success. Record correct outcomes, response time, retries and total cost. Check whether confidence scores match the answers you can verify.
Compare caching and long-context costs ↗