NOUMENOpen app
NOUMEN

Qwen · Open weights

Qwen 3 Coder 480B

Added Mar 17, 2026Open weights

Compare models

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

Loading models…

Use in the app

Qwen 3 Coder 480B, a 480 billion total parameter model with 35B active, and 160 total experts with 8 active. Performs similar to Claude 4 Sonnet in coding benchmarks, but does so at a much lower price.

  • Tools
  • Structured output
Input
Text
Output
Text
Discovery tier
Open weights

Pricing

USD per 1M tokens
Input
$0.13
Output
$0.50

Reference prices as published by the model provider. API example

Specifications

Context window
262K
Max output
65.5K
Average output (7 days)
315 tokens
Longer than 33% of 468 models

Side by side

Qwen 3 Coder 480B 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…

Qwen 3 Coder 480BQwenThis modelGemini 2.5 Flash Preview (09/2025)GoogleRemove
ProviderQwenGoogle
Discovery tierOpen weightsEfficient
AddedMar 17, 2026Sep 25, 2025
Limits
Context window262K1.05MBest
Max output65.54K65.54K
Average output315 tokens—
Pricing, USD per 1M tokens
Input$0.13Best$0.30
Output$0.50Best$2.50
Cache read—$0.03
Performance
Throughput109.8 tok/s—
Time to first token2.1 s—
Request success (24h)95.1%99.7%Best
Benchmark indexes
Intelligence Index11.912.4Best
Math Index39.356.7Best
Capabilities
ReasoningNoYes
ToolsYesYes
VisionNoYes
Structured outputYesYes
PDF inputNoYes
Audio inputNoYes
Reasoning effort—low, medium, high

Public evaluations

Benchmarks

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

  • Qwen 3 Coder 480B
  • Gemini 2.5 Flash Preview (09/2025)

Intelligence Index

  • Qwen 3 Coder 480B11.9
  • Gemini 2.5 Flash Preview (09/2025)12.4

Qwen 3 Coder 480B: Better than 26% of 326 scored models

3.8Catalog models by score55.1

Indexes

  • Math Index

    Composite of the math evaluations

    • Qwen 3 Coder 480B: 39.3
    • Gemini 2.5 Flash Preview (09/2025): 56.7Top

Document reasoning

  • AA-LCRv1.1

    Reasoning across long documents

    • Qwen 3 Coder 480B: 45.7%
    • Gemini 2.5 Flash Preview (09/2025): 60.0%Top

Reasoning and knowledge

  • Humanity’s Last Exam

    Expert-level questions across academic fields

    • Qwen 3 Coder 480B: 4.5%
    • Gemini 2.5 Flash Preview (09/2025): 8.7%Top

  • GPQA Diamond

    Graduate-level science questions

    • Qwen 3 Coder 480B: 61.8%
    • Gemini 2.5 Flash Preview (09/2025): 76.6%Top

  • MMLU-Pro

    Multi-subject knowledge and reasoning

    • Qwen 3 Coder 480B: 78.8%
    • Gemini 2.5 Flash Preview (09/2025): 83.6%Top

  • IFBench

    Following precise output instructions

    • Qwen 3 Coder 480B: 40.5%
    • Gemini 2.5 Flash Preview (09/2025): 43.5%Top

Coding

  • LiveCodeBench

    Recent competitive programming problems

    • Qwen 3 Coder 480B: 58.5%
    • Gemini 2.5 Flash Preview (09/2025): 62.5%Top

Math

  • AIME 2025

    Competition mathematics, 2025 problems

    • Qwen 3 Coder 480B: 39.3%
    • Gemini 2.5 Flash Preview (09/2025): 56.7%Top

  • AIME

    Competition mathematics

    • Qwen 3 Coder 480B: 47.7%
    • Gemini 2.5 Flash Preview (09/2025): —

  • MATH-500

    Competition math problems

    • Qwen 3 Coder 480B: 94.2%
    • Gemini 2.5 Flash Preview (09/2025): —

Legacy evaluations

  • Terminal-Bench Hard

    Earlier terminal task suite

    • Qwen 3 Coder 480B: 18.9%Top
    • Gemini 2.5 Flash Preview (09/2025): 14.4%

  • τ²-Bench Telecom

    Tool use in a customer-support simulation

    • Qwen 3 Coder 480B: 43.6%Top
    • Gemini 2.5 Flash Preview (09/2025): 28.4%

  • AA-LCR (unversioned)

    Long-context reasoning, earlier grading

    • Qwen 3 Coder 480B: 45.7%
    • Gemini 2.5 Flash Preview (09/2025): 60.0%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.

Throughput
109.8 tok/s
Faster than 65% of 253 models
Time to first token
2.1 s
Quicker than 75% of 253 models
Prompt cache hits
75.3%
Request success (24h)
95.1%
Reliable

Cache hits are the share of input tokens recently served from a prompt cache across all requests. They are not a forecast of your own hit rate.

Hourly request success · last 72 hoursLowest hour 52.0%
  • 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": "qwen/qwen3-coder",
    "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/qwen3-coder",
    messages=[{"role": "user", "content": "Explain this stack trace."}],
)
print(reply.choices[0].message.content)

Qwen

Other Qwen models, starting with the same family.