NOUMENOpen app
NOUMEN

Qwen · Open weights

Qwen2.5 VL 72B TEE

Added Jan 1, 2024Open weights

Compare models

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

Loading models…

Use in the app

Qwen2.5 Vision-Language 72B model with multimodal capabilities. Running inside a TEE (Trusted Execution Environment), with provider attestation support.

  • Vision
Input
Text, Images
Output
Text
Discovery tier
Open weights

Pricing

USD per 1M tokens
Input
$0.70
Output
$0.70

Reference prices as published by the model provider. API example

Specifications

Context window
65.5K
Max output
8.2K
Average output (7 days)
84 tokens
Longer than 8% of 468 models

Side by side

Qwen2.5 VL 72B TEE 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 VL 72B TEEQwenThis modelDeepSeek V4 ProDeepSeekRemove
ProviderQwenDeepSeek
Discovery tierOpen weightsOpen weights
AddedJan 1, 2024Apr 24, 2026
Limits
Context window65.54K1.05MBest
Max output8.19K384KBest
Average output84 tokens416 tokens
Pricing, USD per 1M tokens
Input$0.70$0.70
Output$0.70Best$1.40
Cache read—$0.07
Performance
Throughput—56.2 tok/s
Time to first token—2.6 s
Request success (24h)—99.8%
Benchmark indexes
Intelligence Index—36.0
Coding Index—68.8
Capabilities
ReasoningNoYes
ToolsNoYes
VisionYesNo
Structured outputNoYes
Reasoning effort—none, high, max

Public evaluations

Benchmarks

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

  • Qwen2.5 VL 72B TEE
  • DeepSeek V4 Pro

Intelligence Index

  • Qwen2.5 VL 72B TEE—
  • DeepSeek V4 Pro36.0
3.8Catalog models by score55.1

Indexes

  • Coding Index

    Composite of the coding evaluations

    • Qwen2.5 VL 72B TEE: —
    • DeepSeek V4 Pro: 68.8

Agentic work

  • AutomationBench-AA

    Workflow automation, penalized for guardrail violations

    • Qwen2.5 VL 72B TEE: —
    • DeepSeek V4 Pro: 56.7%

  • AutomationBench-AA tasks completed

    Workflows finished end to end without a guardrail violation

    • Qwen2.5 VL 72B TEE: —
    • DeepSeek V4 Pro: 25.9%

  • AA-Briefcasev1.1

    Agentic knowledge work, rated head to head

    • Qwen2.5 VL 72B TEE: —
    • DeepSeek V4 Pro: 1,256 Elo

  • GDPval-AAv2.1

    Economically valuable occupational tasks, rated head to head

    • Qwen2.5 VL 72B TEE: —
    • DeepSeek V4 Pro: 1,449 Elo

Document reasoning

  • GDP.pdf

    Professional reasoning over PDF documents, all-pass rate

    • Qwen2.5 VL 72B TEE: —
    • DeepSeek V4 Pro: 11.4%

  • AA-LCRv1.1

    Reasoning across long documents

    • Qwen2.5 VL 72B TEE: —
    • DeepSeek V4 Pro: 80.3%

Reasoning and knowledge

  • Humanity’s Last Exam

    Expert-level questions across academic fields

    • Qwen2.5 VL 72B TEE: —
    • DeepSeek V4 Pro: 41.0%

  • GPQA Diamond

    Graduate-level science questions

    • Qwen2.5 VL 72B TEE: —
    • DeepSeek V4 Pro: 92.8%

Coding

  • Terminal-Benchv4.0

    Practical software engineering in a terminal

    • Qwen2.5 VL 72B TEE: —
    • DeepSeek V4 Pro: 14.1%

  • SciCode

    Scientific computing in Python

    • Qwen2.5 VL 72B TEE: —
    • DeepSeek V4 Pro: 51.0%

Legacy evaluations

  • AA-LCR (unversioned)

    Long-context reasoning, earlier grading

    • Qwen2.5 VL 72B TEE: —
    • DeepSeek V4 Pro: 80.3%

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": "TEE/qwen2.5-vl-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="TEE/qwen2.5-vl-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.