HyperAI

Main

Home

GPU

Console
Studio
Docs
Pricing

Pulse

News

Resources

Papers
Notebooks
Datasets
Wiki

Benchmarks

SOTA
LLM Models
GPU Leaderboard

Community

Conferences

Utility

Search

Blog Terms of Service Privacy Policy
Acceptable Use Policy About
English

HyperAI

Command Palette

Search for a command to run...

Build the Future of Artificial Intelligence

HyperAI AgentBeta

HyperAI Agent is in beta. Answers may be inaccurate, so check important details.

What can I help with?

Find resources, learn the platform, or manage your workspaces

Kimi K3.1 Pro | LLM Models | HyperAI

HyperAI

Kimi K3.1 Pro

Kimi2026-10-23

At a glance

Intelligence Index: 46.5
Cost per task: $3.68
Time per task: 37m 7s
Output tokens per task: 72.2K
Input / output price per 1M tokens: $2.00 / $10.00
Cached input per 1M tokens: $0.20

Evaluations

Intelligence Index components

AA-Briefcase v1.1
weight
1674
AA-Omniscience v1.1
·
29.9
FrontierCode-AA 2.0
·
25.5%
Terminal-Bench-Science 0.1
34.3%
CritPt v2
49.5%
HLE-Diamond
·
37.3%
Terminal-Bench 4.0
·
67.2%
AutomationBench-AA
73.3%
GDP.pdf
31.4%

Humanity's Last Exam
52.5%
CritPt
24.6%
AA-LCR v1.1: Long Context Reasoning
86.3%
GDPval-AA v2.1

: 2026-10-23


: ,


: Kimi K3.1-Pro Licence

·

1861

SciCode

60.8%

MMMU-Pro

83.9%

MLCR-AA: Medical Long Context Reasoning

38.3%

FrontierCode 1.1

47.1%

AA-AnalystAgent

38.8%

Multimodal Knowledge Agent (MKA)

43.9%

VIALS-AA

45.3%

AA-3D: Interactive 3D Experience

1112

CWE-Bench-AA

60.8%

CyberGym-E2E-AA

94.7%

DeepsecBench-AA

36.7%

15%

General

weight

Coding

weight

Other evaluations

Specifications

Released

Parameters

total, active

Context window

tokens

Input

Text

Image

Output

Reasoning

Weights

Open weights, commercial license

Download weights

Size

Large

Data from Artificial Analysis

Updated

LLM Models

15%

10%

5%

2.8T

104B

1M

2026-11-11