# Llama 4 Maverick

> Meta's flagship natively-multimodal MoE model with 400B total / 17B active parameters

- Maker: Meta
- Line: Llama 4
- Canonical page: https://ai-tldr.dev/models/llama-4-maverick/
- Updated: 2026-06-17

## In short

Llama 4 Maverick is Meta's flagship model in the Llama 4 herd, announced April 5, 2025. It has a 1M-token context window. You can use it through Together AI and Fireworks AI.

## Facts

| Fact | Value | Source |
| --- | --- | --- |
| Context window | 1M | [1] |
| Licence | Llama 4 Community License Agreement · open weights | [1] |
| Released | 5 April 2025 | [2] |
| Parameters | 400B total · 17B active (128 experts) | [1] |
| Knowledge cutoff | August 2024 | [1] |

## Sources

1. [Llama 4 Maverick model card (Hugging Face)](https://huggingface.co/meta-llama/Llama-4-Maverick-17B-128E-Instruct)
2. [The Llama 4 herd announcement](https://ai.meta.com/blog/llama-4-multimodal-intelligence/)

## Overview

Llama 4 Maverick is Meta's flagship model in the Llama 4 herd, announced April 5, 2025. It is a natively multimodal, mixture-of-experts (MoE) model with 17 billion active parameters drawn from 400 billion total parameters across 128 experts, using alternating dense and MoE layers with early-fusion multimodality so text and vision are processed in a single backbone.

It accepts multilingual text and image input and produces multilingual text and code output, with a 1M-token context window. Knowledge cutoff is August 2024. Meta positions it as the best multimodal model in its class, reporting that it beats GPT-4o and Gemini 2.0 Flash on a range of benchmarks while reaching comparable reasoning and coding results to DeepSeek v3.

The model is released under the Llama 4 Community License Agreement and is open-weight, downloadable from Hugging Face and hosted by providers including Together AI and Fireworks AI.

## Benchmarks

- MMLU Pro: 80.5% ([source](https://huggingface.co/meta-llama/Llama-4-Maverick-17B-128E-Instruct))
- GPQA Diamond: 69.8% ([source](https://huggingface.co/meta-llama/Llama-4-Maverick-17B-128E-Instruct))
- MMMU: 73.4% ([source](https://huggingface.co/meta-llama/Llama-4-Maverick-17B-128E-Instruct))
- MMMU Pro: 59.6% ([source](https://huggingface.co/meta-llama/Llama-4-Maverick-17B-128E-Instruct))
- MathVista: 73.7% ([source](https://huggingface.co/meta-llama/Llama-4-Maverick-17B-128E-Instruct))
- ChartQA: 90% ([source](https://huggingface.co/meta-llama/Llama-4-Maverick-17B-128E-Instruct))
- DocVQA (test): 94.4% ([source](https://huggingface.co/meta-llama/Llama-4-Maverick-17B-128E-Instruct))
- LiveCodeBench: 43.4% ([source](https://huggingface.co/meta-llama/Llama-4-Maverick-17B-128E-Instruct))
- MGSM: 92.3% ([source](https://huggingface.co/meta-llama/Llama-4-Maverick-17B-128E-Instruct))

## Strengths

- Native multimodality (text + image) via early fusion in a single model
- Strong multimodal benchmark results — 73.4 MMMU, 90.0 ChartQA, 94.4 DocVQA
- Efficient MoE design: only 17B of 400B parameters active per token
- Long 1M-token context window for large-document and multi-image tasks
- Broad multilingual support across 12 languages

## Best for

- Multimodal assistants combining image understanding with chat
- Long-document and multi-image analysis at up to 1M-token context
- Multilingual text generation and coding
- Self-hosted, open-weight deployments needing frontier multimodal quality

## How to access

- [Together AI](https://www.together.ai/models/llama-4-maverick) — model id `meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8`
- [Fireworks AI](https://fireworks.ai/models/fireworks/llama4-maverick-instruct-basic) — model id `accounts/fireworks/models/llama4-maverick-instruct-basic`

## Llama 4 versions

- Llama 4 Maverick (5 April 2025) — this page
- [Llama 4 Scout](https://ai-tldr.dev/models/llama-4-scout/) (5 April 2025)
- [Llama 4 Behemoth](https://ai-tldr.dev/models/llama-4-behemoth/) (April 2025)

## FAQ

### Is Llama 4 Maverick open-weight?

Yes. Meta releases the weights under the Llama 4 Community License Agreement, and they can be downloaded from Hugging Face (meta-llama/Llama-4-Maverick-17B-128E-Instruct). It is also hosted by third-party providers such as Together AI and Fireworks AI. The license is a custom community license rather than a standard OSI-approved open-source license.

### How many parameters does Llama 4 Maverick have?

Maverick is a mixture-of-experts model with 400 billion total parameters spread across 128 experts, but only 17 billion parameters are active per token. This sparse design gives it large-model quality while keeping inference cost closer to a 17B dense model. It is the higher-expert sibling of Llama 4 Scout (16 experts, 109B total).

### What is the context window of Llama 4 Maverick?

Llama 4 Maverick supports a 1 million-token context window per its Hugging Face model card. Its sibling Llama 4 Scout supports an even larger 10 million-token window. Both accept multilingual text and image input and output multilingual text and code, with a knowledge cutoff of August 2024.

## Links

- [Llama official site](https://www.llama.com/)
- [Llama 4 Maverick model card (Hugging Face)](https://huggingface.co/meta-llama/Llama-4-Maverick-17B-128E-Instruct)
- [The Llama 4 herd announcement](https://ai.meta.com/blog/llama-4-multimodal-intelligence/)
- [Together AI — Llama 4 Maverick API](https://www.together.ai/models/llama-4-maverick)
- [Fireworks AI — Llama 4 Maverick](https://fireworks.ai/models/fireworks/llama4-maverick-instruct-basic)

Source: AI/TLDR LLM registry, https://ai-tldr.dev/models/llama-4-maverick/