Moonshot AI · llm model

Kimi K3.1 Pro

Release Ruler for Kimi K3.1 Pro: released Oct 23, 2026, $2 per 1M input tokens, license not stated. Each thin stem on the ruler is another model we track.

Kimi K3.1 Pro is Moonshot AI's latest language model in our dataset, released Oct 23, 2026. Open weights are promised but not yet published. Moonshot's post-trained successor to Kimi K3, a 3T-class model with a 1M-token context, live in Kimi apps and the API on Oct 23, 2026; open weights are promised by Nov 4. Verified against vendor sources on Oct 26, 2026.

Reported Last verified · All models · JSON · Markdown

  • ReleasedOct 23, 2026
  • Context window1.05M1,048,576 tokens
  • LicenceWeights unpublishedNot yet published

Modalities not verified

Key facts

Kimi K3.1 Pro: key facts with sources
FactValueSource
VendorMoonshot AIn/a
ReleasedOct 23, 2026 Vercel's AI Gateway changelog, dated Oct 23, 2026, lists K3.1 Pro as available that day; Artificial Analysis also dated its launch write-up Oct 23. Moonshot's blog carries no date; Reuters reported the launch on Oct 24 Beijing time.source
Statusgenerally availablen/a
API idkimi-k3.1-pro A faster sibling, kimi-k3.1-pro-highspeed, launched alongside it.source
Context window1,048,576 tokens (1.05M) Moonshot states 1M tokens; the pricing page gives 1,048,576.source
Parameters2.8T total, 104B active, mixture-of-experts Figures are from Artificial Analysis; Moonshot itself only says "3T-class".source
LicenceNot yet published Moonshot calls it an open-weight model and says the full weights will be released by Nov 4, 2026. Until the weights and license are published we do not label it open-weights.source
Pricing$2 in / $10 out per 1M tokens Highspeed variant: $4.00 input, $20.00 output, $0.20 cached input. Rates checked by Moonshot on Oct 24, 2026.source

Where Kimi K3.1 Pro sits in our data

How Kimi K3.1 Pro differs from Kimi K3.1 Pro Highspeed

Read from the recorded facts of the two models, not from a head-to-head test. We have not run either.

  • Listed price is $2 in and $10 out, against $4 in and $20 out for Kimi K3.1 Pro Highspeed (per 1M tokens).
  • Both list a 1.05M-token context window.
  • Both were released on Oct 23, 2026.
  • Facts for the other model: Kimi K3.1 Pro Highspeed.

What the sources say

These are claims from the linked pages, not our measurements.

  • Moonshot reports 65.7% on Terminal-Bench 4.0 (Claude Code harness), against 66.4% for Claude Opus 5.5 and 58.2% for GPT-6 Astra. This is a vendor-run figure.

Notes

  • Artificial Analysis scored it 55 on its Intelligence Index at launch (4th, behind Claude Opus 5.5 at 58, Claude Sonnet 5.5 at 56 and Muse Wonder Full at 56) and 67% on Terminal-Bench 4.0 in its own run, which differs from Moonshot's 65.7%.
  • The faster tier, Kimi K3.1 Pro Highspeed, has its own page: /models/kimi-k3-1-pro-highspeed.

What we tested

Nothing yet. We have not run Kimi K3.1 Pro ourselves, so every fact on this page is Reported from the sources below. When we do test it, the scope will be stated here and in our test logs. See methodology.

Not verified, left out

We could not confirm these on a vendor page, so we do not state them.

  • Input and output modalities were not confirmed on Moonshot's own pages.
  • Weights and license are not published yet; Moonshot says they arrive by Nov 4, 2026.

Sources

More from Moonshot AI

All 88 models