NOUMENOpen app
NOUMEN

Meta · Open weights

Llama 4 Maverick

Added Sep 5, 2025Open weights

Compare models

Choose up to 3 models to compare.0 of 3 chosen

Loading models…

Use in the app

Llama 4 Maverick, a 17 billion active parameter model with 128 experts, is the best multimodal model in its class, beating GPT-4o and Gemini 2.0 Flash across a broad range of widely reported benchmarks, while achieving comparable results to the new DeepSeek v3 on reasoning and coding—at less than half the active parameters. Llama 4 Maverick offers a best-in-class performance to cost ratio with an experimental chat version scoring ELO of 1417 on LMArena.

  • Tools
  • Vision
  • Structured output
Input
Text, Images
Output
Text
Discovery tier
Open weights

Pricing

USD per 1M tokens
Input
$0.19
Output
$0.65
Cache read
$0.05

Reference prices as published by the model provider. API example

Specifications

Context window
1M
Max output
65.5K
Average output (7 days)
114 tokens
Longer than 11% of 473 models

Side by side

Compare models

Put Llama 4 Maverick next to up to three other models. The comparison stays in the link, so you can share it.

Compare models

Choose up to 3 models to compare.0 of 3 chosen

Loading models…

TryGemini 2.5 Flash Lite Preview (09/2025)Qwen3 Coder NextMistral Medium 3.1All 3

Public evaluations

Benchmarks

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

Intelligence Index

9.3

Better than 18% of 326 scored models

3.8Catalog models by score55.1

Indexes

  • Coding Index

    16.3

    Composite of the coding evaluations

    Better than 11% of 168 scored models

  • Math Index

    19.3

    Composite of the math evaluations

    Better than 19% of 118 scored models

Agentic work

  • AutomationBench-AA

    0.2%

    Workflow automation, penalized for guardrail violations

    Better than 3% of 173 scored models

  • AutomationBench-AA tasks completed

    0.0%

    Workflows finished end to end without a guardrail violation

    Better than 0% of 173 scored models

Document reasoning

  • GDP.pdf

    0.8%

    Professional reasoning over PDF documents, all-pass rate

    Better than 2% of 171 scored models

  • AA-LCRv1.1

    50.0%

    Reasoning across long documents

    Better than 29% of 309 scored models

Reasoning and knowledge

  • Humanity’s Last Exam

    4.9%

    Expert-level questions across academic fields

    Better than 16% of 323 scored models

  • GPQA Diamond

    67.1%

    Graduate-level science questions

    Better than 25% of 291 scored models

  • MMLU-Pro

    80.9%

    Multi-subject knowledge and reasoning

    Better than 56% of 132 scored models

  • IFBench

    43.0%

    Following precise output instructions

    Better than 34% of 215 scored models

Coding

  • Terminal-Benchv4.0

    0.0%

    Practical software engineering in a terminal

    Better than 0% of 171 scored models

  • SciCode

    31.7%

    Scientific computing in Python

    Better than 6% of 174 scored models

  • LiveCodeBench

    39.7%

    Recent competitive programming problems

    Better than 30% of 129 scored models

Math

  • AIME 2025

    19.3%

    Competition mathematics, 2025 problems

    Better than 19% of 118 scored models

  • AIME

    39.0%

    Competition mathematics

    Better than 47% of 72 scored models

  • MATH-500

    88.9%

    Competition math problems

    Better than 43% of 73 scored models

Legacy evaluations

  • Terminal-Bench Hard

    6.8%

    Earlier terminal task suite

    Better than 15% of 212 scored models

  • τ²-Bench Telecom

    17.8%

    Tool use in a customer-support simulation

    Better than 7% of 214 scored models

  • AA-LCR (unversioned)

    50.0%

    Long-context reasoning, earlier grading

    Better than 29% of 309 scored models

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.

Prompt cache hits
7.3%
About $0.18 per 1M input tokens with caching
Request success (24h)
99.6%
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 87.5%
  • 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": "meta-llama/llama-4-maverick",
    "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="meta-llama/llama-4-maverick",
    messages=[{"role": "user", "content": "Explain this stack trace."}],
)
print(reply.choices[0].message.content)

Meta

Other Meta models, starting with the same family.