NOUMENOpen app
NOUMEN

Moonshot · Frontier

Kimi Latest

Added May 3, 2026

Compare models

Choose up to 3 models to compare.1 of 3 chosen

Loading models…

Use in the app

Always points to the newest recommended Kimi model, currently Kimi K3.1 Pro. Reasoning is always enabled. Select Kimi K3 explicitly to retain its checkpoint and pricing; K3.1 Pro Highspeed is available separately.

  • Reasoning
  • Tools
  • Vision
  • Structured output
  • Video input
Input
Text, Images, Video
Output
Text
Reasoning effort
low · high · max
Discovery tier
Frontier

Pricing

USD per 1M tokens
Input
$2.00
Output
$10.00
Cache read
$0.20
Cache write
$2.00
$4.00 for a 1-hour cache

Reference prices as published by the model provider. API example

Specifications

Context window
1M
Max output
—

Side by side

Kimi Latest compared

Specifications, prices, speed and benchmark scores next to each other. Benchmarks below show every compared model too.

Compare models

Choose up to 3 models to compare.1 of 3 chosen

Loading models…

Kimi LatestMoonshotThis modelKimi K2.5MoonshotRemove
ProviderMoonshotMoonshot
Discovery tierFrontierOpen weights
AddedMay 3, 2026Jan 26, 2026
Limits
Context window1.05MBest262.14K
Max output—65.54K
Average output—437 tokens
Pricing, USD per 1M tokens
Input$2.00$0.38Best
Output$10.00$2.00Best
Cache read$0.20$0.07Best
Cache write$2.00—
Performance
Throughput59.7 tok/s91.9 tok/sBest
Time to first token5.6 s1.3 sBest
Request success (24h)83.5%99.8%Best
Benchmark indexes
Intelligence Index55.1Best23.5
Coding Index—46.8
Capabilities
ReasoningYesNo
ToolsYesYes
VisionYesYes
Structured outputYesNo
Video inputYesNo
Reasoning effortlow, high, max—

Public evaluations

Benchmarks

Independent evaluation results, with how each score ranks among the models in the Noumen catalog.

  • Kimi Latest
  • Kimi K2.5

Intelligence Index

  • Kimi Latest55.1
  • Kimi K2.523.5

Kimi Latest: Better than 99% of 326 scored models

3.8Catalog models by score55.1

Indexes

  • Coding Index

    Composite of the coding evaluations

    • Kimi Latest: —
    • Kimi K2.5: 46.8

Agentic work

  • AutomationBench-AA

    Workflow automation, penalized for guardrail violations

    • Kimi Latest: 73.3%
    • Kimi K2.5: —

  • AutomationBench-AA tasks completed

    Workflows finished end to end without a guardrail violation

    • Kimi Latest: 46.0%
    • Kimi K2.5: —

  • AA-Briefcasev1.1

    Agentic knowledge work, rated head to head

    • Kimi Latest: 1,678 Elo
    • Kimi K2.5: —

  • GDPval-AAv2.1

    Economically valuable occupational tasks, rated head to head

    • Kimi Latest: 1,846 EloTop
    • Kimi K2.5: 839 Elo

Document reasoning

  • GDP.pdf

    Professional reasoning over PDF documents, all-pass rate

    • Kimi Latest: 31.4%
    • Kimi K2.5: —

  • AA-LCRv1.1

    Reasoning across long documents

    • Kimi Latest: 86.3%Top
    • Kimi K2.5: 78.0%

Reasoning and knowledge

  • Humanity’s Last Exam

    Expert-level questions across academic fields

    • Kimi Latest: 52.5%Top
    • Kimi K2.5: 30.7%

  • GPQA Diamond

    Graduate-level science questions

    • Kimi Latest: —
    • Kimi K2.5: 87.9%

  • IFBench

    Following precise output instructions

    • Kimi Latest: —
    • Kimi K2.5: 70.2%

Coding

  • Terminal-Benchv4.0

    Practical software engineering in a terminal

    • Kimi Latest: 67.2%
    • Kimi K2.5: —

  • SciCode

    Scientific computing in Python

    • Kimi Latest: 60.8%
    • Kimi K2.5: —

Legacy evaluations

  • Terminal-Bench Hard

    Earlier terminal task suite

    • Kimi Latest: —
    • Kimi K2.5: 34.8%

  • τ²-Bench Telecom

    Tool use in a customer-support simulation

    • Kimi Latest: —
    • Kimi K2.5: 95.9%

  • AA-LCR (unversioned)

    Long-context reasoning, earlier grading

    • Kimi Latest: 86.3%Top
    • Kimi K2.5: 78.0%

Scores from Artificial Analysis, updated Oct 26, 2026. Artificial Analysis

Live routing data

Performance and reliability

Recent averages across requests routed for this model. Your own speed depends on prompt length, reasoning effort and load.

Throughput
59.7 tok/s
Faster than 41% of 250 models
Time to first token
5.6 s
Quicker than 29% of 250 models
Prompt cache hits
96.7%
About $0.26 per 1M input tokens with caching
Request success (24h)
83.5%
Unstable

Cache hits are the share of input tokens recently served from a prompt cache across all requests. They are not a forecast of your own hit rate.

Hourly request success · last 72 hoursLowest hour 49.0%
  • Healthy (98% or more)
  • Degraded (90–98%)
  • Down (under 90%)
  • No requests

Developers

Use it through the API

Noumen’s API is OpenAI-compatible: any Chat Completions client works with a base URL and a key.

Read the quickstart
curl https://www.noumen.xyz/api/v1/chat/completions \
  -H "Authorization: Bearer $NOUMEN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "moonshotai/kimi-latest",
    "reasoning_effort": "low",
    "messages": [{ "role": "user", "content": "Explain this stack trace." }]
  }'
import os
from openai import OpenAI

client = OpenAI(base_url="https://www.noumen.xyz/api/v1", api_key=os.environ["NOUMEN_API_KEY"])
reply = client.chat.completions.create(
    model="moonshotai/kimi-latest",
    messages=[{"role": "user", "content": "Explain this stack trace."}],
)
print(reply.choices[0].message.content)

Moonshot

Other Moonshot models, starting with the same family.