NOUMENOpen app
NOUMEN

k2 · Open weights

K2-Think

Added Jul 26, 2025Open weights

Compare models

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

Loading models…

Use in the app

K2-Think is a 32B open-weights general reasoning model with strong competitive math performance. Benchmarks: AIME 2024 90.83, AIME 2025 81.24, GPQA-Diamond 71.08, LiveCodeBench v5 63.97.

Input
Text
Output
Text
Discovery tier
Open weights

Pricing

USD per 1M tokens
Input
$0.17
Output
$0.68

Reference prices as published by the model provider. API example

Specifications

Context window
128K
Max output
32.8K

Side by side

K2-Think 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…

K2-Thinkk2This modelPhi 4 MinimicrosoftRemove
Providerk2microsoft
Discovery tierOpen weightsOpen weights
AddedJul 26, 2025Jul 26, 2025
Limits
Context window128K128K
Max output32.77KBest16.38K
Average output—24 tokens
Pricing, USD per 1M tokens
Input$0.17$0.17
Output$0.68$0.68
Performance
Request success (24h)—100.0%
Benchmark indexes
Intelligence Index—6.3
Coding Index—3.8
Math Index—6.7
Capabilities
Reasoning effort——

Public evaluations

Benchmarks

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

  • K2-Think
  • Phi 4 Mini

Intelligence Index

  • K2-Think—
  • Phi 4 Mini6.3
3.8Catalog models by score55.1

Indexes

  • Coding Index

    Composite of the coding evaluations

    • K2-Think: —
    • Phi 4 Mini: 3.8

  • Math Index

    Composite of the math evaluations

    • K2-Think: —
    • Phi 4 Mini: 6.7

Document reasoning

  • AA-LCRv1.1

    Reasoning across long documents

    • K2-Think: —
    • Phi 4 Mini: 15.3%

Reasoning and knowledge

  • Humanity’s Last Exam

    Expert-level questions across academic fields

    • K2-Think: —
    • Phi 4 Mini: 4.4%

  • GPQA Diamond

    Graduate-level science questions

    • K2-Think: —
    • Phi 4 Mini: 33.1%

  • MMLU-Pro

    Multi-subject knowledge and reasoning

    • K2-Think: —
    • Phi 4 Mini: 46.5%

  • IFBench

    Following precise output instructions

    • K2-Think: —
    • Phi 4 Mini: 21.1%

Coding

  • LiveCodeBench

    Recent competitive programming problems

    • K2-Think: —
    • Phi 4 Mini: 12.6%

Math

  • AIME 2025

    Competition mathematics, 2025 problems

    • K2-Think: —
    • Phi 4 Mini: 6.7%

  • AIME

    Competition mathematics

    • K2-Think: —
    • Phi 4 Mini: 3.0%

  • MATH-500

    Competition math problems

    • K2-Think: —
    • Phi 4 Mini: 69.6%

Legacy evaluations

  • Terminal-Bench Hard

    Earlier terminal task suite

    • K2-Think: —
    • Phi 4 Mini: 0.0%

  • τ²-Bench Telecom

    Tool use in a customer-support simulation

    • K2-Think: —
    • Phi 4 Mini: 8.2%

  • AA-LCR (unversioned)

    Long-context reasoning, earlier grading

    • K2-Think: —
    • Phi 4 Mini: 15.3%

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

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": "LLM360/K2-Think",
    "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="LLM360/K2-Think",
    messages=[{"role": "user", "content": "Explain this stack trace."}],
)
print(reply.choices[0].message.content)