NOUMENOpen app
NOUMEN

Qwen · Open weights

Qwen3 Coder Next

Added Dec 8, 2025Open weights

Compare models

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

Loading models…

Use in the app

Qwen3 Coder Next is an open-weight coding model built on Qwen3-Next-80B-A3B-Base (hybrid attention + MoE). It is agentically trained at scale on executable tasks and environment interaction, delivering strong coding and tool-use performance at lower inference cost. Native 256K context.

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

Pricing

USD per 1M tokens
Input
$0.25
Output
$0.80
Cache read
$0.0798

Reference prices as published by the model provider. API example

Specifications

Context window
262.1K
Max output
65.5K
Average output (7 days)
182 tokens
Longer than 20% of 465 models

Side by side

Qwen3 Coder Next 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 Coder NextQwenThis modelMistral Medium 3.1MistralRemove
ProviderQwenMistral
Discovery tierOpen weightsBalanced
AddedDec 8, 2025Sep 5, 2025
Limits
Context window262.14KBest131.07K
Max output65.54KBest32.77K
Average output182 tokens60 tokens
Pricing, USD per 1M tokens
Input$0.25Best$0.40
Output$0.80Best$2.00
Cache read$0.0798—
Performance
Throughput119.9 tok/s—
Time to first token1.6 s—
Request success (24h)92.2%100.0%Best
Benchmark indexes
Intelligence Index9.29.2
Coding Index36.2Best20.5
Math Index—38.3
Capabilities
ToolsYesNo
Structured outputYesNo
Reasoning effort——

Public evaluations

Benchmarks

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

  • Qwen3 Coder Next
  • Mistral Medium 3.1

Intelligence Index

  • Qwen3 Coder Next9.2
  • Mistral Medium 3.19.2

Qwen3 Coder Next: Better than 17% of 326 scored models

3.8Catalog models by score55.1

Indexes

  • Coding Index

    Composite of the coding evaluations

    • Qwen3 Coder Next: 36.2Top
    • Mistral Medium 3.1: 20.5

  • Math Index

    Composite of the math evaluations

    • Qwen3 Coder Next: —
    • Mistral Medium 3.1: 38.3

Agentic work

  • AutomationBench-AA

    Workflow automation, penalized for guardrail violations

    • Qwen3 Coder Next: 1.1%
    • Mistral Medium 3.1: 4.1%Top

  • AutomationBench-AA tasks completed

    Workflows finished end to end without a guardrail violation

    • Qwen3 Coder Next: 0.0%
    • Mistral Medium 3.1: 0.2%Top

  • AA-Briefcasev1.1

    Agentic knowledge work, rated head to head

    • Qwen3 Coder Next: 404 Elo
    • Mistral Medium 3.1: 530 EloTop

  • GDPval-AAv2.1

    Economically valuable occupational tasks, rated head to head

    • Qwen3 Coder Next: 513 EloTop
    • Mistral Medium 3.1: 385 Elo

Document reasoning

  • GDP.pdf

    Professional reasoning over PDF documents, all-pass rate

    • Qwen3 Coder Next: 2.6%Top
    • Mistral Medium 3.1: 1.2%

  • AA-LCRv1.1

    Reasoning across long documents

    • Qwen3 Coder Next: 47.0%Top
    • Mistral Medium 3.1: 42.7%

  • MLCR-AA

    Long-context reasoning over medical records

    • Qwen3 Coder Next: —
    • Mistral Medium 3.1: 2.8%

Reasoning and knowledge

  • Humanity’s Last Exam

    Expert-level questions across academic fields

    • Qwen3 Coder Next: 10.1%Top
    • Mistral Medium 3.1: 4.7%

  • GPQA Diamond

    Graduate-level science questions

    • Qwen3 Coder Next: 73.7%Top
    • Mistral Medium 3.1: 58.8%

  • MMLU-Pro

    Multi-subject knowledge and reasoning

    • Qwen3 Coder Next: —
    • Mistral Medium 3.1: 68.3%

  • IFBench

    Following precise output instructions

    • Qwen3 Coder Next: 35.2%
    • Mistral Medium 3.1: 39.8%Top

Coding

  • Terminal-Benchv4.0

    Practical software engineering in a terminal

    • Qwen3 Coder Next: 0.0%Top
    • Mistral Medium 3.1: 0.0%Top

  • SciCode

    Scientific computing in Python

    • Qwen3 Coder Next: 36.2%Top
    • Mistral Medium 3.1: 32.4%

  • LiveCodeBench

    Recent competitive programming problems

    • Qwen3 Coder Next: —
    • Mistral Medium 3.1: 40.6%

Math

  • AIME 2025

    Competition mathematics, 2025 problems

    • Qwen3 Coder Next: —
    • Mistral Medium 3.1: 38.3%

Legacy evaluations

  • Terminal-Bench Hard

    Earlier terminal task suite

    • Qwen3 Coder Next: 18.2%Top
    • Mistral Medium 3.1: 10.6%

  • τ²-Bench Telecom

    Tool use in a customer-support simulation

    • Qwen3 Coder Next: 79.5%Top
    • Mistral Medium 3.1: 40.6%

  • AA-LCR (unversioned)

    Long-context reasoning, earlier grading

    • Qwen3 Coder Next: 47.0%Top
    • Mistral Medium 3.1: 42.7%

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
119.9 tok/s
Faster than 68% of 251 models
Time to first token
1.6 s
Quicker than 86% of 251 models
Prompt cache hits
96.8%
About $0.0853 per 1M input tokens with caching
Request success (24h)
92.3%
Unstable

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 75.0%
  • 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-next",
    "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-next",
    messages=[{"role": "user", "content": "Explain this stack trace."}],
)
print(reply.choices[0].message.content)

Qwen

Other Qwen models, starting with the same family.