Kimi K3.1 Pro

Identity and evidence for this model. Performance depends on the endpoint, task and measurement conditions.

Updated: 2026-10-24

Availability & identity

Not in the current trial pool; connect your own endpoint to test.

Moonshot Kimi

Version and API identifier

k3.1-pro

Official source · The pay-as-you-go API uses kimi-k3.1-pro; personal Kimi Coding subscriptions use k3.1-pro. IDs and billing bases differ. The official 1M context and long-task capabilities are not verified by short tasks. Weights are planned by November 4 and were not yet published when checked.

Read the first-look review and original answers

Endpoint pricing

Unknown or not billed per token. No zero-cost assumption is used.

Full price table

Standard test evidence

No published standard report yet. Task suitability has not been established.

All models · Method and limitations