NOUMENOpen app
NOUMEN

Qwen · Open weights

Qwen2.5 72B

Added Jul 3, 2025Open weights

Compare models

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

Loading models…

Use in the app

Great multilingual support, strong at mathematics and coding, supports roleplay and chatbots.

  • Tools
Input
Text
Output
Text
Discovery tier
Open weights

Pricing

USD per 1M tokens
Input
$0.36
Output
$0.40

Reference prices as published by the model provider. API example

Specifications

Context window
131.1K
Max output
8.2K
Average output (7 days)
166 tokens
Longer than 16% of 468 models

Side by side

Qwen2.5 72B 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…

Qwen2.5 72BQwenThis modelQwen 3.8 27B AntislopQwenRemove
ProviderQwenQwen
Discovery tierOpen weightsOpen weights
AddedJul 3, 2025Oct 5, 2026
Limits
Context window131.07K262.14KBest
Max output8.19K32.77KBest
Average output166 tokens299 tokens
Pricing, USD per 1M tokens
Input$0.36$0.10Best
Output$0.40Best$0.60
Cache read—$0.05
Performance
Throughput—33.8 tok/s
Time to first token—2.1 s
Request success (24h)—99.1%
Benchmark indexes
Intelligence Index7.7—
Math Index14.0—
Capabilities
ReasoningNoYes
ToolsYesYes
VisionNoYes
Reasoning effort——

Public evaluations

Benchmarks

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

  • Qwen2.5 72B
  • Qwen 3.8 27B Antislop

Intelligence Index

  • Qwen2.5 72B7.7
  • Qwen 3.8 27B Antislop—

Qwen2.5 72B: Better than 11% of 326 scored models

3.8Catalog models by score55.1

Indexes

  • Math Index

    Composite of the math evaluations

    • Qwen2.5 72B: 14.0
    • Qwen 3.8 27B Antislop: —

Reasoning and knowledge

  • Humanity’s Last Exam

    Expert-level questions across academic fields

    • Qwen2.5 72B: 3.6%
    • Qwen 3.8 27B Antislop: —

  • GPQA Diamond

    Graduate-level science questions

    • Qwen2.5 72B: 49.1%
    • Qwen 3.8 27B Antislop: —

  • MMLU-Pro

    Multi-subject knowledge and reasoning

    • Qwen2.5 72B: 72.0%
    • Qwen 3.8 27B Antislop: —

  • IFBench

    Following precise output instructions

    • Qwen2.5 72B: 36.9%
    • Qwen 3.8 27B Antislop: —

Coding

  • LiveCodeBench

    Recent competitive programming problems

    • Qwen2.5 72B: 27.6%
    • Qwen 3.8 27B Antislop: —

Math

  • AIME 2025

    Competition mathematics, 2025 problems

    • Qwen2.5 72B: 14.0%
    • Qwen 3.8 27B Antislop: —

  • AIME

    Competition mathematics

    • Qwen2.5 72B: 16.0%
    • Qwen 3.8 27B Antislop: —

  • MATH-500

    Competition math problems

    • Qwen2.5 72B: 85.8%
    • Qwen 3.8 27B Antislop: —

Legacy evaluations

  • Terminal-Bench Hard

    Earlier terminal task suite

    • Qwen2.5 72B: 4.5%
    • Qwen 3.8 27B Antislop: —

  • τ²-Bench Telecom

    Tool use in a customer-support simulation

    • Qwen2.5 72B: 34.5%
    • Qwen 3.8 27B Antislop: —

Scores from Artificial Analysis, updated Oct 26, 2026.

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": "qwen/qwen-2.5-72b-instruct",
    "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="qwen/qwen-2.5-72b-instruct",
    messages=[{"role": "user", "content": "Explain this stack trace."}],
)
print(reply.choices[0].message.content)

Qwen

Other Qwen models, starting with the same family.