AI news · Kimi
Kimi K3.1 Pro: Highspeed, open weights and document tasks
What Kimi K3.1 Pro adds, how Highspeed differs, when open weights are expected and a practical document-analysis prompt you can check yourself.
Kimi K3.1 Pro is the new release to watch for people working with long documents, code and 3D projects. Moonshot’s official announcement says Pro and Pro Highspeed are available in Kimi’s products and API. Availability in another app is a separate question.
At a glance
- Context window
- 1,048,576 tokens
Sources: Kimi K3.1 Pro API guide
- Access
- kimi-k3.1-pro / kimi-k3.1-pro-highspeed
Sources: Kimi K3.1 Pro API guide
What is new in Kimi K3.1 Pro?
Moonshot describes a model built for long work across coding, document research and 3D creation, with a context window of roughly one million tokens. The launch includes Kimi K3.1 Pro and a Highspeed option. Features inside Kimi Work or Kimi Code are product workflows; a model in a different chat app does not automatically inherit those tools.
Sources: Moonshot: Kimi K3.1 Pro · Kimi K3.1 Pro API guide
Kimi K3.1 Pro Highspeed and API access
The official API guide lists kimi-k3.1-pro and kimi-k3.1-pro-highspeed. Moonshot claims Highspeed is two to three times faster; we have not independently measured that. API usage and a Kimi subscription are separate purchases. Check the current official rate card and the selected model before starting a large document job.
Sources: Moonshot: Kimi K3.1 Pro · Kimi K3.1 Pro API guide · Kimi API pricing
Are Kimi K3.1 Pro open weights available?
The announcement promises the complete weights by November 4. On October 27, that is a future commitment, not evidence that you can download them today. Check the official release, licence and hardware requirements when they appear. Open weights also do not mean free hosted API usage. Avoid replacing a working workflow solely on the basis of a promised release.
Sources: Moonshot: Kimi K3.1 Pro
Try this document-research prompt first
A long context window does not guarantee a faithful summary. Test a brief containing a firm requirement, a proposal and an unresolved decision. The correct result below keeps two entrances as required, Saturday as proposed and the roof material undecided. Ask for supporting quotes, then inspect every row. For a 3D project, resolve these questions before requesting a model; an attractive render can still violate the brief.
Analyze only this fictional pavilion brief: The pavilion must have two entrances. Saturday was proposed for the review, but no date was approved. Roof material is undecided. Return a table of requirements, proposals and open questions with an exact supporting quote for each. Do not invent dimensions, materials, costs or approval. Finish with the three questions needed before creating a 3D model.Questions, answered
Does a one-million-token context guarantee accuracy?
No. It describes input capacity, not guaranteed retrieval or correct interpretation. Use quotes and an answer key to check the important details.
Can I generate the pictured pavilion with this prompt?
The cover is editorial illustration, not Kimi output. This prompt prepares a brief; creating or exporting a 3D model needs a suitable tool and further instructions.
Can I try this exact model in GO AI?
Availability of this exact model in GO AI is not confirmed. The links open GO AI with its currently offered models, not a guaranteed selection of the model discussed here.
Sources and editorial note
Reported facts are linked to their sources. Sections marked GO AI analysis contain our interpretation; prompts are suggestions, not recorded test results.