Model board / Kimi K3.1 Pro

Kimi K3.1 Pro

latest Moonshot · Kimi

Verified 2026-10-29 against the pages listed below. Prices are USD per 1M tokens, list rate — no committed-use, enterprise or volume discount.

Input
$2.00
Cached input
$0.2
Cache write
$2.00
Output
$10.00
Context window
1,048,576
Max output
—
Weights
closed
Caching
automatic

flat across the full window — the guide says there is no tiering by context length. Cache write $2.00 at the default 5-minute TTL, $4.00 for 1 hour. A Highspeed variant, kimi-k3.1-pro-highspeed, bills $4.00/$0.20/$20.00 with writes at $4.00/$8.00 for 2–3x the speed — this row is the standard tier

1,048,576 for all K3-series models on the pricing page; max_completion_tokens defaults to 131,072 and can be raised to the full window

cache hit at 10% of input

Watch out

The weights are not published yet — Moonshot's guide says they will be released by Nov 4, and until they are on Hugging Face under a named licence this is a closed row. Reasoning is always on and defaults to max effort, billed as output. Batch is not offered. Moonshot claims coding and agentic gains over K3; this board has run no evaluation of either

What we would use it for

Kimi K3's flat 1M window at two-thirds of its price

What this has cost, week by week

Output price at each refresh since 2026-10-29. A vendor change is the provider changing its price. Our correction means this board changed what it prints and the provider's charge did not move — the two are never merged, because publishing the second as the first would be a price rise that never happened.

10-29 $10.00

Nothing has moved since we started tracking it.

Where these numbers came from

No number on this page was written that was not read from one of those pages on 2026-10-29. If one is wrong, send us the source URL and it is corrected in the next weekly refresh, with a changelog line saying what was wrong — contact@naderu.com.

Jobs this is one of our picks for

Discussed in Naderu Weekly

The rest of the board

This row sits alongside 73 others on the model board, refreshed every week. The same data as JSON, and the terms for using it, are at /board-data.