Ollama
One-command local LLM runtime.
OPEN SOURCECPU-CAPABLE
Meta
Llama 4 Maverick
Meta's open workhorse — huge ecosystem, runs everywhere.
Open weightstextimageREL 2025-04-05
Compare this model →Verified 2026-07-14
Overview
Llama 4 Maverick is Meta's mainstream open-weights MoE model with vision input and a 1M context window, priced at $0.15/$0.60 per 1M via API. Its real superpower is the ecosystem: every local runner, fine-tune toolkit and deployment stack supports Llama first.
Benchmark scores
Methodology →Intelligence18
Codingn/a
Reasoningn/a
Mathn/a
Sources and verification
- Intelligence: Artificial Analysis Intelligence Index · checked 2026-06-12
- Catalogue and pricing: OpenRouter model record · checked 2026-07-14
Strengths and weaknesses
Strengths
- Biggest open ecosystem
- Vision + 1M context
- Very cheap
Trade-offs
- 16K max output
- Older than Chinese open rivals (Apr 2025)
Reach for this model when
Teams standardising on the Llama ecosystem for fine-tuning and local deployment.
Run something like it locally
First-class support in Ollama, LM Studio, llama.cpp and vLLM — quantised builds run on 24GB+ GPUs.