NOUMENOpen app
NOUMEN

Qwen · Balanced

Qwen3 VL 235B A22B Instruct Original

Added Sep 25, 2025

Compare models

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

Loading models…

Use in the app

Routed across multiple providers, which may include Alibaba, a Chinese entity; provider-specific privacy and logging guarantees may vary. Qwen3 Vision‑Language model (235B MoE, ≈22B active) tuned for instruction following and grounded visual QA. Excels at image understanding, dense OCR, charts and diagrams, and multi‑image context. Use this variant when you want concise, direct answers grounded in the visuals.

  • Vision
Input
Text, Images
Output
Text
Discovery tier
Balanced

Pricing

USD per 1M tokens
Input
$0.30
Output
$1.90

Reference prices as published by the model provider. API example

Specifications

Context window
—
Max output
32.8K
Average output (7 days)
288 tokens
Longer than 29% of 468 models

Side by side

Qwen3 VL 235B A22B Instruct Original 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…

Qwen3 VL 235B A22B Instruct OriginalQwenThis modelGemini 2.5 FlashGoogleRemove
ProviderQwenGoogle
Discovery tierBalancedEfficient
AddedSep 25, 2025Jun 5, 2025
Limits
Context window—1.05M
Max output32.77K65.54KBest
Average output288 tokens368 tokens
Pricing, USD per 1M tokens
Input$0.30$0.30
Output$1.90Best$2.50
Cache read—$0.03
Performance
Throughput—197.0 tok/s
Time to first token—7.6 s
Request success (24h)100.0%Best99.7%
Benchmark indexes
Intelligence Index9.99.9
Math Index70.7Best60.3
Capabilities
ReasoningNoYes
ToolsNoYes
VisionYesYes
Audio inputNoYes
Video inputNoYes
Reasoning effort—low, medium, high

Public evaluations

Benchmarks

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

  • Qwen3 VL 235B A22B Instruct Original
  • Gemini 2.5 Flash

Intelligence Index

  • Qwen3 VL 235B A22B Instruct Original9.9
  • Gemini 2.5 Flash9.9

Qwen3 VL 235B A22B Instruct Original: Better than 20% of 326 scored models

3.8Catalog models by score55.1

Indexes

  • Math Index

    Composite of the math evaluations

    • Qwen3 VL 235B A22B Instruct Original: 70.7Top
    • Gemini 2.5 Flash: 60.3

Document reasoning

  • AA-LCRv1.1

    Reasoning across long documents

    • Qwen3 VL 235B A22B Instruct Original: 32.7%
    • Gemini 2.5 Flash: 49.9%Top

Reasoning and knowledge

  • Humanity’s Last Exam

    Expert-level questions across academic fields

    • Qwen3 VL 235B A22B Instruct Original: 6.6%Top
    • Gemini 2.5 Flash: 4.7%

  • GPQA Diamond

    Graduate-level science questions

    • Qwen3 VL 235B A22B Instruct Original: 71.2%Top
    • Gemini 2.5 Flash: 68.3%

  • MMLU-Pro

    Multi-subject knowledge and reasoning

    • Qwen3 VL 235B A22B Instruct Original: 82.3%Top
    • Gemini 2.5 Flash: 80.9%

  • IFBench

    Following precise output instructions

    • Qwen3 VL 235B A22B Instruct Original: 42.7%Top
    • Gemini 2.5 Flash: 39.0%

Coding

  • LiveCodeBench

    Recent competitive programming problems

    • Qwen3 VL 235B A22B Instruct Original: 59.4%Top
    • Gemini 2.5 Flash: 49.5%

Math

  • AIME 2025

    Competition mathematics, 2025 problems

    • Qwen3 VL 235B A22B Instruct Original: 70.7%Top
    • Gemini 2.5 Flash: 60.3%

  • AIME

    Competition mathematics

    • Qwen3 VL 235B A22B Instruct Original: —
    • Gemini 2.5 Flash: 50.0%

  • MATH-500

    Competition math problems

    • Qwen3 VL 235B A22B Instruct Original: —
    • Gemini 2.5 Flash: 93.2%

Legacy evaluations

  • Terminal-Bench Hard

    Earlier terminal task suite

    • Qwen3 VL 235B A22B Instruct Original: 6.8%
    • Gemini 2.5 Flash: 12.1%Top

  • τ²-Bench Telecom

    Tool use in a customer-support simulation

    • Qwen3 VL 235B A22B Instruct Original: 35.1%Top
    • Gemini 2.5 Flash: 14.9%

  • AA-LCR (unversioned)

    Long-context reasoning, earlier grading

    • Qwen3 VL 235B A22B Instruct Original: 32.7%
    • Gemini 2.5 Flash: 49.9%Top

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

Live routing data

Performance and reliability

Recent averages across requests routed for this model. Your own speed depends on prompt length, reasoning effort and load.

Request success (24h)
100.0%
Reliable
Hourly request success · last 72 hoursLowest hour 100.0%
  • Healthy (98% or more)
  • Degraded (90–98%)
  • Down (under 90%)
  • No requests

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": "qwen3-vl-235b-a22b-instruct-original",
    "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="qwen3-vl-235b-a22b-instruct-original",
    messages=[{"role": "user", "content": "Explain this stack trace."}],
)
print(reply.choices[0].message.content)

Qwen

Other Qwen models, starting with the same family.