Llama 4 Maverick

Meta · Released Apr 5, 2025

Open SourcevisionreasoningbudgetCompare this model

$0.15

Input price / MTok

#4 of 27

$0.6

Output price / MTok

#4 of 27

Context window

#14 of 28

118

Tokens / sec

#9 of 21

69.8

GPQA Diamond

#18 of 19

73.4

MMMU

#6 of 8

Summary

Llama 4 Maverick is a open-source language model from Meta, released on Apr 5, 2025.

At $0.15 input and $0.6 output per million tokens, it ranks #4 of 27 for input price, placing it among the most affordable options.

Its context window of 1M tokens ranks #14 of 28 for context depth in this dataset.

It targets vision, reasoning, budget workloads.

Throughput is measured at 118 tokens per second (ranked #9 of 21 for throughput).

Flagship of the Llama 4 herd; stronger reasoning and vision grounding than Scout.

Benchmark	Score	Max	%	Rank	Source
GPQA Diamond	69.8	100	69.8%	#18/19	Meta AI (self-reported on llama.com)
MMMU	73.4	100	73.4%	#6/8	Meta AI (self-reported on llama.com)

Models with the most shared benchmark coverage and closest scores to Llama 4 Maverick.

Llama 4 Maverick

Llama 4 Scout

o4-mini

Claude Sonnet 4.6

Gemini 2.5 Pro

GPQA Diamond#18 of 19

Graduate-level science reasoning

MMMU#6 of 8

Multimodal understanding across 30 subjects

Compare Llama 4 Maverick side by side

Pre-filled with 5 similar models