Moonshot AI's Batch API now supports the kimi-k3.1-pro and kimi-k3 models on both CN and Global platforms, previously limited to kimi-k2.7-code and kimi-k2.6. Batch inference remains priced at 60% of standard rates, but jobs on the new models also incur Cache Write fees, and cache hits are best-effort due to asynchronous, concurrent processing.
Read the full story on platform.kimi.ai
What to doReview the Batch API Guide and BatchJob pricing before running kimi-k3.1-pro or kimi-k3 batch jobs, since Cache Write fees now apply.
- Provider
- Moonshot AI
- Source
- platform.kimi.ai