← Models

Llama 4 Maverick

Meta · Llama 4

Meta's larger open-weight Llama 4 MoE model for multimodal assistants and controlled deployments.

Part of Llama family · Other versions: Llama 4 Scout
Type
multimodal
Context
1M tokens
Max Output
33K tokens
Status
current
API Access
Yes
License
Llama Community
open-weights self-hosted reasoning enterprise customization moe vision
Released April 2025 · Updated May 16, 2026

Overview

Freshness note: Model capabilities, deployment options, and licensing terms can change. This profile is a point-in-time snapshot last verified on May 16, 2026.

Llama 4 Maverick is Meta’s larger released Llama 4 model for teams that prioritize deployment control and model customization. Meta’s official launch material describes Maverick as a natively multimodal MoE model with 17B active parameters, 400B total parameters, 128 experts, and a 1 million token context window.

Capabilities

Maverick-class open models are often used for internal assistants, controlled-domain reasoning, multimodal workflows, and custom tool-enabled systems. Meta positions Maverick as the product-workhorse Llama 4 route for general assistant and chat use cases, image understanding, creative writing, and coding/reasoning workloads where an open-weight deployment is required.

Technical Details

Meta’s official materials list multilingual text and image input, multilingual text and code output, and native multimodal pretraining with early fusion. Maverick is the higher-capability released sibling to Scout in Meta’s Llama 4 family, but real-world quality still depends heavily on serving stack, quantization choices, context implementation, and evaluation discipline.

Pricing & Access

Access can come through self-hosted infrastructure or cloud providers exposing compatible endpoints. Cost structure differs substantially from closed APIs because infrastructure and operations become major factors.

Best Use Cases

Good fit for regulated environments, on-prem or private-cloud assistants, and teams that want deeper control over model lifecycle and deployment economics.

Comparisons

Compared with GPT-5.5 and Claude Opus 4.7, Maverick offers more deployment control but usually requires more engineering effort for equivalent polish. Compared with Llama 4 Scout, Maverick is the higher-quality route with heavier serving needs. Compared with Qwen3.6-27B, choice depends on language needs, license fit, and serving strategy.