Routeway Logo
Back to Models
MoonshotAI

Kimi K2 0905

Available

Description

Kimi K2 0905 is the September update of Kimi K2 0711. It is a large-scale Mixture-of-Experts (MoE) language model developed by Moonshot AI, featuring 1 trillion total parameters with 32 billion active per forward pass. It supports long-context inference up to 256k tokens, extended from the previous 128k. This update improves agentic coding with higher accuracy and better generalization across scaffolds, and enhances frontend coding with more aesthetic and functional outputs for web, 3D, and related tasks. Kimi K2 is optimized for agentic capabilities, including advanced tool use, reasoning, and code synthesis. It excels across coding (LiveCodeBench, SWE-bench), reasoning (ZebraLogic, GPQA), and tool-use (Tau2, AceBench) benchmarks. The model is trained with a novel stack incorporating the MuonClip optimizer for stable large-scale MoE training.

At a Glance

Key pricing and model details available for this model.

Input price

$0.39

per 1M tokens

Output price

$1.90

per 1M tokens

Context window

262K

tokens

Hallucination rate

17.9%

Token Pricing

Token pricing normalized to per-million-token rates.

Input / 1M tokens

$0.39

Output / 1M tokens

$1.90

Cache Read / 1M tokens

$0.20

Token Pricing Details

Rates are shown per 1M tokens for easier comparison.

Input / 1M tokens$0.39
Input unit1M tokens
Output / 1M tokens$1.90
Output unit1M tokens
Cache Read / 1M tokens$0.20
Cache Read unit1M tokens

Feature Availability

Capabilities explicitly listed in the current payload.

LLM

Available

Yes

Vision

Available

Yes

Function calling

Available

Yes

Reasoning

Available

Yes

Supported Parameters

frequency_penalty
logit_bias
max_completion_tokens
presence_penalty
reasoning_effort
response_format
stop
temperature
tool_choice
tools
top_p

Artificial Analysis

Index scores currently reported for this model.

Intelligence Index

30.9

Coding Index

25.9

Math Index

57.3

Category Radar

Aggregated from the benchmark values present for reasoning, code, math, and accuracy.

Benchmark Breakdown

Detailed benchmark results drawn from the current payload.

Intelligence Index

index

Overall 'how smart' score for an AI, combining reasoning, math, coding, and knowledge.

30.9

Reported score

Coding Index

index

How well the model handles real programming tasks.

25.9

Reported score

Math Index

index

Composite score measuring mathematical reasoning and problem-solving.

57.3

Reported score

MMLU-Pro

reasoning

A broad and difficult knowledge-and-reasoning benchmark across many subjects.

81.9%

Reported score

GPQA

reasoning

Graduate-level science questions designed to be difficult to shortcut.

76.7%

Reported score

HLE

reasoning

A very hard expert-level exam across a wide range of subjects.

6.3%

Reported score

LiveCodeBench

code

Fresh programming tasks meant to test current coding ability.

61%

Reported score

SciCode

code

Coding tasks drawn from real scientific workflows.

30.7%

Reported score

AIME 2025

math

The 2025 AIME benchmark used to reduce data leakage concerns.

57.3%

Reported score

IFBench

accuracy

Measures how precisely the model follows detailed instructions.

41.7%

Reported score

LCR

accuracy

Tests long-context reasoning over large documents and conversations.

52.3%

Reported score

TerminalBench Hard

code

A harder coding-agent benchmark for complex multi-step terminal tasks.

23.5%

Reported score

Tau2

accuracy

Evaluates realistic agent behavior in tool-using support workflows.

73.4%

Reported score

Code Samples

Quick start with the Routeway API

import OpenAI from 'openai';

const openai = new OpenAI({
  baseURL: "https://api.routeway.ai/v1",
  apiKey: "<YOUR_API_KEY>",
});

async function main() {
  const completion = await openai.chat.completions.create({
    model: "kimi-k2-0905",
    messages: [
      {
        role: "user",
        content: "Explain quantum computing in simple terms"
      }
    ]
  });

  console.log(completion.choices[0].message);
}

main();