Llama 4 Maverick 17B-128E
MoEToolsVisionMeta · 400B (17B active) · Mixture of Experts
Multimodal MoE with 128 experts — 17B active, 1M context
5.9K
468
2025-04
1M tokens
Use cases
Your hardware
16 GB Reference Desktop
Quantization options
What is quantization?| Quant | Bits | VRAM | Disk | Quality | Status |
|---|---|---|---|---|---|
| Q2_K | 2 | 128.6 GB | 122.2 GB | low | Too heavy · ~1 t/s |
| Q3_K_M | 3 | 179.8 GB | 171.1 GB | moderate | Too heavy · ~1 t/s |
| Q4_K_M | 4 | 205.4 GB | 195.6 GB | good | Too heavy · ~1 t/s |
| Q5_K_M | 5 | 256.6 GB | 244.5 GB | good | Too heavy · ~1 t/s |
| Q6_K | 6 | 307.8 GB | 293.4 GB | excellent | Too heavy · ~1 t/s |
| Q8_0 | 8 | 410.3 GB | 391.2 GB | excellent | Too heavy · ~1 t/s |
| F16 | 16 | 820.1 GB | 782.3 GB | lossless | Too heavy · ~1 t/s |
About this model
Llama 4 Maverick 17B-128E is a 400B sparse mixture-of-experts (MoE) model activating 17B active per token from Meta in the Llama family, designed primarily for chat, vision, reasoning, code workflows. Supports structured tool calling; use Q4_K_M or higher when strict JSON schema adherence matters. Supports up to 1M context tokens; start at 4K–8K with Q8_0 KV cache on tight GPUs.
Hardware planning baselines start at 223.5 GB minimum memory and 372.5 GB recommended memory. At Q4_K_M, weights plus baseline runtime overhead use roughly 205.4 GB of RAM/VRAM and 195.6 GB of disk space. At 400B, Llama 4 Maverick 17B-128E belongs firmly in the workstation and multi-GPU tier. Running it comfortably requires 372.5 GB of total memory across dual RTX GPUs, high-memory Apple Silicon (Mac Studio Ultra/Max), or MoE hybrid CPU-RAM offloading—mobile devices and standard 16 GB laptops are too constrained for practical speeds.
Deploy Llama 4 Maverick 17B-128E locally via Ollama (`llama4:maverick`), Ollama with Flash Attention enabled. Under Llama 4 Community License Agreement, review official weight cards before commercial redistribution.
Highlights
- Local Fit Baseline223.5 GB minimum and 372.5 GB recommended RAM/VRAM.
- Primary QuantQ4_K_M targets 205.4 GB memory and 195.6 GB disk.
- Context & Capabilities1M max context · Sparse MoE · Tool calling.
System requirements by device
Memory planning targets 223.5 GB minimum RAM/VRAM and 372.5 GB recommended memory. The baseline quant is Q4_K_M, estimated at 205.4 GB memory and 195.6 GB disk. Actual runtime usage shifts with context length, KV cache precision, and OS background overhead.
| Device | Notes |
|---|---|
| Mac (Apple silicon) | Requires a 48 GB–192 GB Mac Studio or MacBook Pro Max/Ultra with enough unified wired budget for 205.4 GB. |
| Windows / Linux GPU | Use dual/multi-GPU CUDA splitting (--tensor-split) or MoE CPU expert offloading (-ot) paired with 32–128 GB DDR5 system RAM. |
| Phones & Compact Edge Boards | Not practical on phones (223.5 GB min RAM exceeds Android LMK and iOS Jetsam limits). Pick a 2B–4B model on mobile. |
Start with a short session and leave memory for the operating system. A file fitting on disk does not guarantee that it will load in memory.
Estimated memory by quantization
These are rough estimates, not tested minimum requirements. Check the actual file and runtime documentation; context, offloading and operating-system overhead can change the memory needed.
| Quant Tier | Estimated memory | Disk Size | Quality |
|---|---|---|---|
| Q2_K | 128.6 GB | 122.2 GB | low |
| Q3_K_M | 179.8 GB | 171.1 GB | moderate |
| Q4_K_M | 205.4 GB | 195.6 GB | good |
| Q5_K_M | 256.6 GB | 244.5 GB | good |
| Q6_K | 307.8 GB | 293.4 GB | excellent |
| Q8_0 | 410.3 GB | 391.2 GB | excellent |
| F16 | 820.1 GB | 782.3 GB | lossless |
How to run Llama 4 Maverick 17B-128E on your device
Start with a short prompt and the smallest quant that still gives the quality you need. A clean first run is more useful than loading a larger file with no memory headroom.
Run Llama 4 Maverick 17B-128E on Multi-GPU or High-RAM Desktop
- Plan for 223.5 GB minimum combined memory and 372.5 GB recommended memory for Q4_K_M.
- Use llama.cpp `-ot ".ffn_.*_exps.=CPU"` or `--tensor-split` across multiple GPUs so attention and KV cache remain in fast VRAM.
- Enable Flash Attention (`-fa on`) and Q8_0 KV cache (`-ctk q8_0 -ctv q8_0`) to prevent context buffers from exceeding GPU headroom.
- Avoid attempting to run workstation-scale weights on smartphones or 8 GB laptops; use a 2B–8B checkpoint on portable devices.
Run Llama 4 Maverick 17B-128E on High-Memory Apple Silicon Mac
- Verify you are on a 48 GB–192 GB Mac Studio or MacBook Pro Max/Ultra capable of holding 205.4 GB in wired unified memory.
- Launch via `llama-server` or LM Studio with MLX/Metal enabled and set initial context conservatively to 4K–8K.
- Monitor Activity Monitor Memory Pressure to confirm zero SSD swap thrashing during token generation.
- Step down to a lighter quantization (e.g., Q3_K_M or IQ4_XS) if total memory utilization exceeds 75%.
Llama 4 Maverick 17B-128E vs alternatives
| Model | Params | Context | Min RAM | License |
|---|---|---|---|---|
| Llama 4 Maverick 17B-128E | 400B (17B active) | 1M | 223.5 GB | Llama 4 Community |
| Llama 3.1 405B | 405B | 128K | 226.3 GB | Llama 3.1 Community |
| Llama 4 Scout 17B | 109B (17B active) | 128K | 60.9 GB | Llama 4 Community |
| Llama 3.3 70B | 70B | 128K | 39.1 GB | Llama 3.3 Community |
| Llama 3.2 11B Vision | 11B | 128K | 6.1 GB | Llama 3.2 Community |
| Llama 3.1 8B | 8B | 128K | 4.5 GB | Llama 3.1 Community |
Use this table to choose the smallest nearby model that still matches the job. A related model with lower RAM can be the better local choice when the current page fits only at the edge of available memory.
Frequently asked questions
- What is Llama 4 Maverick 17B-128E?
- Llama 4 Maverick 17B-128E is a 400B sparse mixture-of-experts model (17B active active per token) from Meta, supporting up to 1M tokens with 223.5 GB minimum RAM and 372.5 GB recommended RAM.
- How much VRAM or RAM does Llama 4 Maverick 17B-128E require?
- At Q4_K_M, Llama 4 Maverick 17B-128E occupies roughly 205.4 GB of active memory and 195.6 GB on disk. Add 2–6 GB of spare headroom for KV cache buffers, runtime scratch space, and OS apps (372.5 GB recommended).
- What is the best way to run Llama 4 Maverick 17B-128E locally?
- Use `ollama run llama4:maverick` for a quick CLI/API setup, or load Q4_K_M in LM Studio or llama.cpp with Flash Attention enabled when you want explicit KV cache and layer control.
- Can Llama 4 Maverick 17B-128E run on a consumer laptop or smartphone?
- No smartphone or 8–16 GB laptop can hold Llama 4 Maverick 17B-128E cleanly in local memory. It targets multi-GPU NVIDIA workstations, high-RAM Mac Studios, or MoE CPU-RAM hybrid rigs with at least 223.5 GB total memory.
- How does Llama 4 Maverick 17B-128E compare to Llama 3.1 405B?
- Llama 3.1 405B (405B, 226.3 GB min RAM) sits closest in size. Compare active parameters, context ceiling (1M vs 128K), and license terms (Llama 4 Community) before switching.